EP0264315A1 - Wave propagation structures for the suppression of over-voltages and the absorption of transitory waves - Google Patents

Abstract

Quadrupole / tripole wave propagation structures, such as lines, cables and electronic components, comprising a non-linear lossy dielectric, absorbing both overvoltages (varistor effect in the time domain) and fast transients (filter effect low pass, in the frequency domain). Application to protections against lightning strike, nuclear electromagnetic pulse, electrostatic discharges and RFI in general.

Description

Today electronic components, having a nonlinear electrical behavior are well known: Thus silicon carbide surge arresters (SiC) and resistances to zinc oxide (ZnO) are commonly used to absorb parasitic overvoltages, along high-voltage lines, or in low-voltage electrical circuits and electronic circuits.

The electrical characteristic of such a component is described, approximately, by a relation of the type

where 1 is the current flowing through it, under an applied voltage U, n being the coefficient of non-linearity, describing "the slope" of non-linearity (typically varying from 3 to 10 for a resistance to SiC, and from 20 to 70 for resistance to ZnO) and k a constant defining the range of conductivity obtained.

It is important, in practice, to be able to have such characteristics, not only with solid materials (sintered SiC and ZnO), but also with composite materials, based on plastics, polymers etc. thermoplastics and thermosets, to facilitate the production by low temperature techniques (compression, injection, extrusion etc., rolling) of parts in the process and / or in quantity, or even when flexibility is necessary, such as for use for electric wires and cables.

Such composites have been described in the literature, with SiC (Brevets France no 1,260,453 and 1,363,222), with ZnO (French patent no 2,547,451) and with various other metal oxides (US patent no 1,246,829 ). The non-linearity coefficients obtained are in the range from 3 to 5. More recently, the Chomerics Company has marketed a flexible composite based on silicon carbide and titanium ("CHOTRAP") with a non-linearity coefficient of l 'order of 7, on a range of currents from 3 to 4 decades.

The applications of the composites described were essentially for the production of sleeves for medium and high-voltage insulated cable terminations: indeed, a low current created by the equipotentials of the electric field, causes a gradient distribution of this field, more favorable, avoiding snapping at the tips. (French Patents no 1,194,221, 1,260,453, 1,363,222; France Patent Application no 2,423,036). It is essentially a "localized" layer in which there is no propagation effect, and permanently conducting a weak current, to define the new equidistributed equipotentials. It is an object of this patent to describe wave propagation structures (in opposition to a non-linear dipole component), in which the non-linear medium is incorporated along the direction of propagation of a wave. electromagnetic (characterizing the disturbance by overvoltage), as

dielectric.

That is to say, the non-linear medium intervenes in

the distributed electrical elements of the structure.

It is another goal, to describe such a structure,

which due to its distribution does not present parasitic effects (parasitic inductance, parasitic capacitance), characterizing the components with dipolar structure.

It is another goal to describe such a structure,

where the nonlinear medium is not requested when the applied electric voltage is normal: in other words this dielectric acts normally, like conventional insulator. Only in the event of disturbing overvoltages appearing, this dielectric conducts and "short-circuits" these overvoltages (to ground, or to another conductor).

It is another object to describe such a structure, in which the dielectric constant, as well as the dielectric losses of this non-linear dielectric increase with the overvoltage. More particularly the distributed (lossy) capacity increases significantly, introducing a low-pass filter effect (RC line), and a characteristic impedance change

of the structure, and the corresponding reflections of the electromagnetic waves.

It is another object to describe such a structure, in which such non-linear effects are obtained with a dielectric and magnetic composite, that is to say where magnetic effects are grafted onto the characteristics described above. with magnetic losses, as described for example in France patent no 78.33.385.

It is an ultimate aim to describe such a structure, in which the effects of suppressing overvoltage (suppression in the time domain) are combined with the effects of filtering by dielectric and / or magnetic absorption and by reflection (in the frequency domain).

Such a structure therefore acts both as a distributed voltage limiter and as a distributed low-pass filter (insofar as the losses increase with frequency).

The practical interest of the invention lies in the concept of a suppression of distributed overvoltage, making it possible to distribute the dissipated power (overvoltages conducted to ground or to another conductor) and to eliminate the drawbacks of the components (dipoles) conventional nonlinear, such as unfavorable responses in fast transient conditions.

Finally, the distribution of non-linear effects, in cable-type structures, introduces the new concept of integrating protection functions in electrical power or information transfer links, that is to say in elements where these overvoltages and transients

are generated and transmitted.

Interest in the protection of energy distribution or signal and information distribution type connections, against lightning strikes, against the electromagnetic pulse of a nuclear explosion (HEMP), against electrostatic discharges, against overvoltages generated in a system (SEMP), automobile distribution network, with rupture of inductive loads, is obvious. The invention will be described in detail, using a number of exemplary embodiments, using a number of figures.

• L représente l'inductance série distribuée, comportant la"self interne", due au conducteur et la "self externe", due à la présence éventuelle d'un milieu magnétique entourant le conducteur ;
• R représente les pertes liées à cette inductance série L, telles que par exemple l'effet peau du conducteur, et, le cas échéant, les pertes magnétiques
  
• C représente la capacité shunt due au diélectrique séparant le conducteur "chaud" de la masse (ou autre conducteur) ;
• r représente la résistance équivalent série des pertes du diélectrique, où l' angle de pertes
  

Thus, FIG. 1 represents the well-known diagram of a conventional quadrupole (tripole) propagation structure, with its distributed elements:

• L represents the distributed series inductor, comprising the "internal self", due to the conductor and the "external self", due to the possible presence of a magnetic medium surrounding the conductor;
• R represents the losses linked to this series inductor L, such as for example the skin effect of the conductor, and, where appropriate, the magnetic losses
  
• C represents the shunt capacity due to the dielectric separating the "hot" conductor from the mass (or other conductor);
• r represents the equivalent series resistance of the dielectric losses, where the angle of losses
  

According to the invention, the dielectric used is represented by a non-linear composite, that is to say behaving essentially as an insulator at normal voltages (applied to the quadrupole (tripole), but becoming essentially conductive, when overvoltages appear at the terminals. quadrupole (tripole).

• FIG. 2 represents the elementary distributed diagram of the structure according to the invention, in which a variable conductivity G (U) is introduced by increasing conduction with the voltage U applied, and in which C (U) and r (U) describe the permittivity and the increasing dielectric losses with this same voltage U.
• FIG. 3 represents the elementary distributed diagram of another structure according to the invention, in which a conventional insulator (that is to say independent of the voltage) is interposed between the non-linear dielectric and the ground (or other conductor ), introducing a capacity C ', considered without loss.
• FIG. 4 represents the application of the principles according to the invention to a coaxial structure (such as a line element, a cable etc.), in which the non-linear dielectric is placed between the "hot" conductor and a sheath or external braid.
• FIG. 5 represents the application of the principles according to the invention to a multiconductor cable, with protection against overvoltages of common mode and of symmetrical mode.
• FIG. 6 represents the application of the principles according to the invention to a flat line element or a multicore ribbon cable.
• FIG. 7 represents the application of the principles, according to the invention, to an insulated medium or high-voltage cable, comprising a discontinuity in one of the electrodes.
• FIG. 8 represents the application of the principles according to the invention to a medium or high-voltage cable, with protection by distribution of the gradient of the electric field, combined with a low-pass filter function.
• FIG. 9 represents the application of the principles, according to the invention, to a case of free wave propagation (plane wave, guided or non-guided).
• FIG. 10 represents the application of the principles, according to the invention, to a quadrupole (tripole) capacity.

In order to explain the invention, the detailed description of these different forms of implementation will follow, with, progressively, the description of the non-linear dielectrics used: it goes without saying that these examples are not limiting , where, in particular, the non-linear dielectrics described can be applied without distinction to the various embodiments, as the non-linear dielectric media can be compact, or made of thermoplastic or thermosetting composites. The examples chosen are typical; it goes without saying also that the principles used can be applied to all other structures with free or guided wave propagation.

In FIG. 4, which describes a typical coaxial cable, according to the invention, 1 represents the central conductor, made of conductive material (metal or alloy), solid, stranded, or in layers. This conductor can possibly be covered with a thin conductive layer, chemically compatible with the non-linear dielectric layer 2. This last layer, in fact, must be in direct contact of good quality, so as not to introduce parasitic conduction effects. (speaking on G).

For the same reasons, the external conductor 4 must be compatible, and ensure good contact with the dielectric 2: In the figure, a layer 3 can be provided for this purpose, consisting, for example, of a metallization of the dielectric 2 (Silvering , Indium deposition, Conductive polymer, Metallic thin foil, Colamine, etc.).

Layer 4 can be a conventional metal braid, a sheet of conductive wires, a metal tube, etc., intended to make the electrical connection. Layer 5 represents a mechanical protection of the line element or the cable, such as a layer of plastic, polymer, armor, etc. It can also, or additionally represent a more or less conductive, more or less absorbent dielectric layer (according to the techniques described in French patent no. 78.333.35), or even be produced with a non-linear dielectric of the same kind as that of layer 2, to suppress currents / overvoltages of external common mode.

The medium 2 is produced, in a first example, by a flexible non-linear dielectric, produced according to known means, described in the references in the preambles.

For example, a mixture of a sintered SiC powder, with high conductivity (NORTON 254 type, from Carborundum) by n or p doping, crushed to a grain size (multi-crystalline aggregates, comprising multiple active interfaces) in the range of 30 to 200 µ, optionally selected in particle size distribution, is integrated, by mixing, in a flexible matrix material, such as plastic (PVC, etc.) or polymer (Silicone, EPDM etc.) with a SiC concentration by volume from 15% to 75%, equivalent to about 20 _^0/0 to 94% by weight, depending on the density of the matrix material.

The choice of the exact powder (doping) and its concentration in the mixture are directly linked to the normal operating voltage, for a given dielectric thickness (voltage for which the dielectric is essentially insulating) and to what is considered "overvoltage" in the use of the cable (for which the dielectric will be essentially conductive by the appearance of Tunnel effect and / or Schottky effect (2 diodes polarized in opposite direction) "short-circuiting" the barrier of potention, due to the interface insulating.

The mixture is then extruded / injected, and crosslinked, if necessary, around the central conductor. As a numerical example, the details of such a mixture will be given below, making it possible to obtain a coefficient of non-linearity of 3 to 4, covering 3 to 4 decades of current. A charge at 73% by weight of doped SiC, with a regular particle size distribution between 50 and 150 μ, is extruded under the diameter of 5 mm, around a central conductor of diameter 2 mm, and then the external electrode put in place. For this embodiment, for a cable length of 1 m, a shunt current of 0.72 uA is raised under a voltage U of 100v, increasing to 2.68mA under a voltage U of 1000v. This corresponds to an average non-linearity coefficient of 3.57.

For the voltage of 100v, it is considered, in this context that the dielectric of the cable is essentially insulating; while under 1000v (overvoltage) the dielectric is essentially conductive. The relative permittivity ε of this dielectric (at low voltage) is 12.2 (at 100 Hz), decreasing slightly at increasing frequency: s = 11.0 (at 1 MHz) and 9.1 (at 100 MHz). Its dielectric loss angle (at low voltage) is tgδe = 0.030 (at 1 kHz), 0.015 (at 10 kHz) increasing towards a maximum of approximately 0.021 (at 100 kHz), and decreasing towards 0.011 (at 100 MHz) , typical variations for a composite with intergranular phases.

At increasing voltage, the permittivity as well as the dielectric losses increase. Thus the dielectric constant and the angle of losses tgδe (due to r (U)) are more than tenfold, the dielectric losses equivalent to non-linear conduction (due to G (U)) obviously increasing much faster. We will return in detail, below, to these unexpected effects and characteristics of the invention.

One of the characteristics of the invention consists in the distribution of the Joule effects, in the case of significant overvoltages, of appreciable duration: therefore, much higher powers can be allowed, compared with the conventional SiC and ZnO protection components. Particular effects may appear in the case of composite dielectrics, according to the invention.

So when the power dissipated (by Joule effect) in such a dielectric heats up the non-linear composite significantly, it tends to expand, and the contacts between particles tend to decrease, causing an additional effect of resistance with coefficient of positive temperature (PTC) .This effect, in the case of a longer overvoltage, causes self-limiting of the current, ie automatic protection, obviously at the expense of the suppressing effect of overvoltage.

It is also obvious, that if the mechanical realization of the matrix of the non-linear composite does not allow this expansion (for example external sheath 4,5 very rigid, compact non-linear dielectric, or even composite with thermosetting matrix) the effect PTC cannot play, or even, a negative temperature coefficient (NTC) effect can appear, by the increase in pressure on the particles in the matrix. (Effect which also exists in the event of the appearance of pressure from other origins).

The application of an overvoltage to such a type of propagation element will obviously cause the entire overvoltage to appear at the input of the structure, a waveform whose amplitude will decrease (as well as the dissipation), as it spreads along. In order to provide for a better distribution of the load, as well as of the dielectric stress, it is possible to increase the thickness of the non-linear dielectric towards the ends of the line, or else to apply to it a less conductive non-linear composite.

In a second example, applied to a structure similar to that of FIG. 4, we consider a non-linear dielectric based on agglomerates of zinc oxide crystals, as used in varistors (MOV's) (electronic components protection). These agglomerates are obtained by crushing sintered parts, for example, and placed in a matrix material as described above, with a mass load of at least 30% of agglomerates, containing at least half of grains of dimensions greater than 100 µ. A low charge (a few%) of conductive graphite is added to promote the number of contacts between agglomerates.

The exponent of non-linearity obtained is approximately 5.1, with conductivities (coefficient k) of an order of magnitude greater.

These two examples make it possible to generalize, with regard to the basic constituent, that is to say the agglomerates of conductive crystals separated by interstices which have little or no conductivity (between conductive crystals), causing the non-linear effects.

Many other polycrystalline structures (than those cited) are usable and have been described in the scientific and technical literature: We cite here other oxides such as alumina, magnesia, titanium and bismuth oxide and., other carbides, such as that of titanium, bismuth, boron etc. ; barium titanates (used in thermistors), sulfides, such as Zinc sulfide (used in electroluminescent panels), ferroelectric compounds (used in barrier coatings), Polycrystalline silicon, etc. We can also synthesize such structures, as we will see below.

As a third example, adapted to currents of an order of magnitude still beyond (ie even lower operating voltages) a non-linear dielectric, using agglomerates of crystals of SiC and titanium carbide , with additions of very fine conductive particles, intended to perfect the contacts between agglomerates. This composite is known commercially as CHOTRAP (Chomerics), already mentioned. It makes it possible to reach exponants of non-linearity n of the order of 7.

The relative permittivity of this composite is approximately 15 (100 Hz), decreasing to 13 (1 MHz), at low test voltage.

The current, for 1 m of cable described, is 3.7 A at a maximum voltage U of 120v (essentially conductive condition). Under the normal operating voltage of 24v, the current is 47 µA in insulating condition.

The dielectric constant in the event of overvoltage is multiplied by a coefficient of 50 to 100, and the same is true for intrinsic dielectric losses (not due to conduction).

Another important aspect of the invention is thus arrived at by the very large variation in the distributed capacity of the structure to completely change the propagation characteristic. Indeed, a line normally adapted (in the event of low voltage) becomes highly mismatched and most of the signal corresponding to the overvoltage is reflected at the input, a third effect which contributes to protection against overvoltage. Immediate applications of these phenomena, by analogy with the TR and ATR cells of radars, are possible. We also mention the fact that the very strong increase in dielectric losses (in the event of overvoltage) increases dielectric absorption, the fourth protective effect.

• - un shuntage résistif (diélectrique essentiellement conducteur) ;
• - une augmentation importante de l'angle de pertes diélectriques (absorbant l'onde em) ;
• - une multiplication importante de la capacité distribuée (accumulant l'énergie sous forme de charge) ;
• - une désadaptation du quadripôle (s'il était adapté avant).

Finally, according to the invention, the following protection / absorption effects appear, in the event of an overvoltage:

• - resistive shunting (essentially conductive dielectric);
• - a significant increase in the angle of dielectric losses (absorbing the em wave);
• - a significant increase in distributed capacity (accumulating energy in the form of charge);
• - a mismatch of the quadrupole (if it was adapted before).

As a fourth example, a non-linear dielectric medium will be indicated, with additional magnetic characteristics, as well as some typical examples of the constitution of such a composite, describing one of the important characteristics of the invention.

The fact of operating at high frequencies, the medium must be crossed by em fields, that is to say its skin effect must be reduced: this condition calls for the use of fine ferromagnetic powders or ferrimagnetic materials.

To obtain high absorption at HF frequencies, these ferrites are a good choice, starting from high permeability, reduced conduction (in the compact state) and various other criteria defined in detail in French Patent No. 78.333.85. Constituted by poorly conductive crystals (characteristics at the base of the ferrites produced), they comprise crystalline interstices, with relatively conductive phases: in other words the classical ferrites are not suitable a priori for achieving non-linear effects.

Nevertheless, ad hoc ferrites can be produced, in which the interstices are optimized for the application according to the invention: for this, it is first necessary to introduce a certain conductivity in the crystals themselves, so as to be able to obtain good conduction under a high electric field (where the Tunnel and / or Schottky effect eliminates the effect of the insulating gaps). It is then necessary to introduce, by weak additions of metals or metallic salts, which do not integrate into the magnetic structure of crystals (or domains), but which segregate in the form of compounds with little or no conductivity, in the interstices between crystals. In this way, "depletion layers", similar to those of non-linear SiC and ZnO, are created by the formation of deserted intergranular phases, with the salts alone of these additions, or in combination with salt components. metallic (Fe, Mn, Zn, Ni, Mg etc.) ferrites.

A second technique according to the invention consists in producing afterwards (after the sintering of the ferrite) such interstices by an ad hoc treatment of the ferrite, which is put into powder form. The application of such essentially insulating thin layers in the form of a deposit in adhesion primer solution, of photopolymerization in the vapor phase of polymer, of chemical treatment of grains, including oxidation in air, etc., is known in itself, in processing techniques in the chemical industry.

An "artificial ferrite" of the first type, combining magnetic effects, and high HF absorption, with the effects of non-linear conduction, was produced, according to the invention. It is of the Mn-Zn type (see cited patent), with the general formula Fe ₂ O ₃ (MnO _x ZnOy) + (SnO, Ti0 ₂ , CaO, Bi ₂ 0 ₃ , CuO, etc.) _z

in which z corresponds to typical impurities (O to some olo), intended to favor the formation of the deserted intergranular phases containing these oxides, (and others, such as the salts of Sb, Pr, Ba, Sr, Nd, Rb , Zr, Co, etc.) and more complex phases, also containing the basic constituents of ferrite (such as ZnO, for example). Generally these additives, composed of oxides strongly basic conductors, with insulating acid oxides favor the formation of interstitial layers with little or no conductivity, and the known factors intervening on the size of the ferrite crystals (such as CaO, for example) make it possible to define, the number of Schottky junctions , ie the voltage where the nonlinear effects take place.

In this general formula x corresponds to the amount of Mno, typically in the range of a mole percentage of 20 to 50% and y to the amount of Zno in the range of a mole percentage of 0 to 40%. (The sum of the percentages x + y + z can be different from the unit insofar as the composition of the ferrite is not stoechrometic, more particularly to meet the condition of good conductivity of the grains). Excess iron Excess iron (appearing in the two valences) is also a factor favoring non-linear phenomena.

Such typical artificial ferrite exemplary thereafter contains a mole percentage of 40% MnO and 14% ZnO (25% and 10 _^0/0 by weight) with 2 mol% Ti0 ₂ and 0.6 mol% of Co, as additives, for the formation of the deserted interstitial layers. Ferrite provides high permeability, with high magnetic losses, for use as an HF absorbent; its dielectric losses show a maximum in HF, characteristic of the Maxwell-Wagner effect, due to a conductive phase and a quasi-insulating phase.

This ferrite can be used in compact form, or in composite, according to the invention. As an example of this last application, a composite produced, with 85% by weight of ferrite (crushed, with a distribution of linear particle size, between 50 and 200 μ in size of agglomerates), and 15% by weight of polyvinyl chloride ( PVC), a conductivity close to that of the cited example of the SiC composite is obtained, with an exponent of non-linearity n = 4 to 4.5 for 4 decades of current.

The electric cable, with the dimensions indicated, for 1 m in length, leads in shunt 0.6 mA at 200v (essentially insulating condition) and 0.45 A at 1000v (essentially conductive condition).

An "artificial ferrite" of the second type uses, for example, an Mn-Zn ferrite powder of the conventional type, (without special interfacial layer). This powder is surface treated in an aqueous or alcoholic solution of silane (such as for example vinyl-tri (β-methoxy-etoxy-silane) ("A-172"), then dried with heat, and annealed to 150 "for 2 hours. The resulting material is then compacted in a press or ground to be integrated into a plastic or polymer (composite) matrix.

As already mentioned, the number of active gaps (depleted areas) determines the overall voltage drop (for a given thickness of non-linear material), and the use of larger, more conductive grains, makes it possible to increase the non-linear current, for a given electric field.

A commercial product, H7C4 power ferrite (TDK), using additions of Si0 ₂ and CaO in the interstices, with an adequate heat treatment (in a controlled atmosphere) to have large crystals, is suitable for increasing the conductivity of an order of magnitude, compared to the example above.

To further increase the conductivity, finally mixtures using these types of ferrites (large crystals, conductive crystals etc.) with additions of agglomerates of SiC and / or ZnO with large grains, and fine metallic additions, possibly ferromagnetic (carbonyl iron , coprecipitated iron-nickel alloys, etc.) may be suitable.

• - une caractéristique magnétique (augmentant L, par l'augmentation de la selfinductance externe) ;
• - des pertes magnétiques, permettant la réalisation de structures à propagation absorbantes.

Such "artificial ferrites" (compact or composite) add to the four effects above

• - a magnetic characteristic (increasing L, by increasing the external selfinductance);
• - magnetic losses, allowing the realization of absorbent propagation structures.

Such magnetic structures, with non-linear conduction are particularly suitable for the suppression of parasites, by clipping (in the time domain) and by HF absorption (in the frequency domain): they represent an ideal solution for interference suppression and immunization EMC where high overvoltages can appear (EMP, lightning strike, inductive cutoff on automobile network), with fast wave fronts (HEMP), with possible transients of very short duration (Corona, automobile ignition parasites , electrostatic discharges).

In FIG. 5, one of the multiple variants of electric cable that can be produced is described, with the principles described with the aid of FIG. 4. It is a multiconductor cable (typically a low electric power distribution cable). -tension), with two phases and earth, or with 3 phases.

The meaning of the numerical references is the same as for FIG. 4: the non-linear dielectric 2 intervenes, between phases, for a double thickness, that is to say a double voltage (voltage between phases), where it occurs in simple , with respect to the earth: The metallic / metallized layer 3 defines the electric field, which determines the non-linear current due to the differential overvoltages. Here again, a common mode protection can be provided by the external layer 5.

In Figure 6, which describes a flat line, as it can be used in flat cables, hybrid circuits, propagation structures on printed circuits, or surface mounted components (SMD).

The conductors 1 and 3 define the electrodes of the line; the non-linear dielectric 2 can use any of the compact or composite materials according to the invention.

It is obvious that the principles of the invention are applicable to still other structures, not described in detail, structures with guided and unguided propagation, using meandering, helical conductors, etc.

Two examples will still be described, using figures, in this sense.

In FIG. 7, representing an insulated cable, 1 corresponds to the bare central conductor, 2 to the "insulating" dielectric, 3 'and 3 "to the localized presence of a ground.

The field gradient, for the mass 3 'can be high, with the peak effect existing at point 7: this point constitutes a weakened place, where the breakdown of the dielectric 2 (normal insulator) can preferably take place.

According to the invention, the dielectric 2 is made completely (or partially, in the radial direction of the cable) by a non-linear dielectric: any local high field gradient, will give rise to a weak conduction, and consequently, to a spreading field lines, avoiding breakdown.

A typical application case corresponds to that of the ignition cable of an automobile engine. In this case, the conductor 1 can be straight (made of conductive or resistant metal, or of a more or less conductive composite, such as a carbon composite) or even be constituted by a helix around a core non-magnetic or magnetic absorbent, according to France patents no.78,333.85 and 86,000,617.

The thickness of the dielectric 2, in conventional insulation is limited in practice (for example, by standards, or by questions of cost price), the point of weakness of the insulation is located at the place 7 where the ignition cable is near or touches a metal part, such as an accelerator control cable, engine part, etc. : the fact of the embodiment according to the invention reduces the safety margin of the insulation.

In the same way, for the case of a very thin conductor 1 (straight or helical), the peak effect specific to this conductor can be "spread" by a non-linear dielectric according to the invention. For example, a conventional absorbent ignition wire on the market uses a helix with a diameter of 2.5 mm, made of metal wire with a diameter of 0.10 mm, with a pitch of about 30 turns per centimeter. An internal layer (or the total layer) in non-linear dielectric, this time reduces the internal electrical stress.

Such ignition cables have electrical connections (attached to wire 1) at the ends, with insulating caps 6, made of rubber, plastic or polymer, to protect the connection with the spark plug or with the distributor.

In the presence of a localized mass 3 ", the same phenomena occur, increased by the fact that the end piece 6 can form a tiny air space with the insulator 2, causing an additional high potential gradient, all of this favoring the breakdown of insulators.

In the latter case, with part or all of the end-piece 6, produced with the non-linear dielectric, according to the invention, it is possible to protect against breakdowns, without the need for very efficient insulating materials, or realizations / delicate assemblies.

It is obvious, as demonstrated in these last examples, that the non-linear function can be combined with the magnetic absorption effect (by the use of a non-linear magnetic compound / composite), giving the structure l absorbent low pass filter effect. A typical combination consists in using a central core (for the helix) with nonlinear magnetic dielectric (protecting against the gradients of the fine wire of the helix), and an external "insulating" layer using one of the nonlinear composites little conductors described (magnetic or non-magnetic), suitable for the high voltages involved.

Another example is shown in Figure 8, to show the possibility of combining such discontinuity protection effects (with their local high potential gradients), with a low-pass absorption filter effect: in this case, too , only part of the cable is covered by the absorbing non-linear magnetic dielectric, case corresponding to the diagram in Figure 2.

In the left part of FIG. 8, the nonlinear magnetic dilectric 2 "is superimposed concentrically with a conventional dielectric 2 '. In this" coaxial "part, the low-pass absorbing effect is joined to that of the non-linearity, but it is also obvious, that its non-linear effects can only intervene for variations in the voltage-overvoltage. (Diagram of Figure 3).

In the right part of FIG. 8, the nonlinear magnetic dielectric 2 "plays the classic role of" equipotential sleeve "(references cited in the preamble), it will cover the stripped part (of the connection), with therefore a configuration similar to that of the previous example. As before, also, the non-linear magnetic dielectric can fill the entire cross section of the dielectric, where we therefore return to an embodiment according to the diagram in FIG. 1.

In FIG. 9, the application of the principles of the invention to a free wave propagation structure is shown.

A plane em wave 8 is incident on the metal surface 3 of a mechanical structure, such as an airplane, vessel etc., this wave coming for example from a Radar beam and the surface 4 being an object to be detected, or else representing a lure.

In the known techniques RAM (absorbent Radar materials), an absorbent dielectric layer 2 is applied to the surface of the object 3, the characteristics of which are chosen to absorb the wave as well as possible (in its crossing of the layer 2) and to reflect this wave as little as possible, directly. The embodiments of such dielectric layers are known, and generally use absorbent magnetic materials.

According to the invention, the dielectric layer 2 is made in part or completely of non-linear absorbent magnetic material, of ad-hoc characteristic impedance.

Under the incidence of sufficient power or by the application of a voltage (by an electrode 9, transparent for waves), the reflectivity of the structure can be controlled between total absorption (disappearance of the object vis with respect to RADAR, this is at minimum RCS) and almost total reflection (maximum RCS), analogous to the mismatch effects for cables, described above.

Finally, in FIG. 10, another application of the invention to a component for electronic circuits is shown, in which the concentration of the capacitive, absorption and non-linearity effects makes it possible to obtain useful practical filter components.

In this figure, the conductor 1 can be made to include the terminal solder connections of the hot conductor, through. It can also be represented by an internal metallization of the passage in the dielectric 2, allowing the passage of a wire, a connection plug, etc.

The dielectric 2 is made of non-linear magnetic absorbent material, as described above ("artificial ferrite"). This gives the effect of a low-pass absorption filter, with shunt conduction protection, under the effect of an overvoltage. A possible metallization 3 ensures contact with the ground electrode 4, forming the quadrupole / tri-pole.

Such components, without the effect of the capacitive filter and the low-pass absorbent filter, have been described recently in the literature, produced in compact ZnO (10th International Aerospace and Ground Conference on Lightning and Static Electricity; Wolff and Earle: "A new form of transient suppressor ", p. 293ff).

Obviously, the component according to the invention can be produced in a multiple form, that is to say where the parallel conductors 1 form a filter connector or else where metallized holes 1, as described, provide a filter base for integrated circuits, connectors, the connections of which then pass through the filter base.

Claims (18)

1) Wave propagation structure, - in which the electromagnetic wave passes through a dielectric, - in which this dielectric is distributed in the direction of wave propagation - in which this dielectric has a non-linear ohmic behavior, in the sense that it is essentially non-conductive at normal operating voltages of this structure, and that it becomes essentially conductive for abnormal electrical overvoltages, - in which this dielectric is a polycrystalline material, comprising interstitial thin layers, causing an electrical effect of the tunnel or Schottky type, under the influence of the high value of the electric field, caused by these overvoltages. 2) Structure according to 1) in which the polycrystalline material is a solid, non-magnetic body, with relatively conductive crystals, separated by thin intergranular layers, essentially insulating, such as for example those existing in crystalline silicon, metal oxides such as l 'Zinc oxide (used in varistors), aluminum oxide, magnesium, titanium, Bismuth etc. ; silicon carbide (used in surge arresters), titanium, boron etc. ; barium titanate, strontium (used in thermostances); ferroelectric compounds (based on Barium and Strontium etc., used in barrier coatings) and polythene-mica, zinc sulfide compounds (used in electroluminescent panels, etc., or, for example, essentially conductive powders treated at the surface to present such essentially insulating layers. 3) Structure according to 1) and 2), in which the polycrystalline material is placed in the form of grains (multi-crystalline aggregates) inside an insulating or poorly conductive matrix material, rigid or flexible, with a sufficient concentration to ensure a at least partial contact between the different grains in this composite, with, if necessary, small additions of conductive grains, to optimize the conductivity. 4) Structure according to 1) to 3) in which this dielectric with non-linear behavior has a dielectric constant and an angle of dielectric losses which increase with the applied voltage and thus introduce a distributed capacity and an increasing dielectric absorption. 5) Structure according to 1) to 4) in which any overvoltage propagating in the structure is partially removed by conduction to ground (action in the time domain), partially stored and absorbed by the increase in capacity to ground, and the increase in dielectric losses (action in the frequency domain), and partially reflected towards the source, by the mismatch of the structure, due to this increase in distributed capacity. 6) Structure according to 1) to 5) in which this nonlinear dielectric consists of a dielectromagnetic solid polycrystalline material, consisting of agglomerates of relatively conductive magnetic crystals, comprising essentially intergranular thin layers essentially insulating, such as for example those of certain ferrimagnetic ceramics , more particularly comprising specific impurities favoring the creation of such intergranular layers, with essentially insulating phases, and, optionally small additions, according to 2) to optimize the non-linearity, or again for example essentially conductive ferrimagnetic powder treated on the surface to present such essentially insulating layers. 7) Structure according to 1), 3) to 6) in which the magnetic polycrystalline material is placed in the form of grains (multi-crystalline aggregates) inside an insulating or poorly conductive matrix material, rigid or flexible, with sufficient concentration to ensure contact at least partially between the different grains in this composite, with, if necessary, small additions of conductive grains, to optimize the conductivity. 8) Structure according to 1). 6) and 7) in which this nonlinear dielectric has a dielectric constant and an angle of dielectric losses which increase with the applied voltage and thus introduce a distributed capacity and an increasing dielectric absorption, as well as a permeability and an angle of losses magnetic independent of the voltage, but the latter increasing with the frequency. 9) Structure according to 1), 6) to 8) in which any propagating overvoltage is partially removed by conduction to ground, partially stored and absorbed by the increase in capacitance to ground and the increase in dielectric losses , partially reflected towards the source, by the mismatch of the structure due to this increase in distributed capacity, and partially absorbed by the magnetic permeability and the magnetic losses (low-pass filter effect). 10) Structure according to 1) to 9) in which all wave propagation is protected, by the suppression, at the same time, of the parasitic voltages exceeding a certain amplitude threshold, as well as the suppression of fast waveforms ( transient), exceeding a certain frequency, in their spectrum, and this independently of their amplitude. 11) Structure according to 1) to 10) in which this non-linear dielectric represents a coefficient of non-linearity equal to or greater than 2, in order to be able to represent, depending on the case, an ad-hoc approximation of the essentially insulating behavior for voltages nominal operating conditions, and essentially conductive behavior for annoying overvoltages. 12) Structure according to 1) to 11) in which at least one conventional insulating layer (vacuum, air, plastic, polymer, ceramic, etc.) is interposed with the non-linear dielectric. 13) Structure according to 1) to 12) represented by a line, line element, or electric cable, of concentric, flat, helical shape, etc., with at least two conductors, or at least one conductor and a mass. - _'14) Structure according to 1) to 12) wherein the non-linear dielectric is crossed by a free propagation wave (plane wave) or a guided wave (wave guide) and the non-linear effect can be controlled . 15) Structure according to 1) to 13) in which the line, line element or electric cable comprises a normal insulator inside and the non-linear magnetic dielectric absorbing outside, and in which a certain length of this dielectric is used as a sleeve for distributing the field gradient, at the ends of this structure. 16) Structure according to 1) to 15) in which the cable is an ignition harness cable for internal combustion engines, with insulating caps, in which part or all of the cable insulation and / or the caps are constituted by non-linear dielectric, thus reducing the dielectric stresses at the points of contact or near a ground, or even between turns of the cable, in the case of a helical conductor. 17) Structure according to 1) to 16) in which the structure represents an electronic inductive or capacitive protection component, quadrupole (tripole), or even an assembly combining at least one inductance and capacitance, at least one of which consists of a component quadrupole (tripole), for making more complex filters. 18) Structure according to 1) to 17) in which the control of the possible expansion of the non-linear dielectric, under the thermal load of an overvoltage, makes it possible to obtain effects of positive temperature coefficients, (with an effect of self-protection), zero, or negative.

Images