FR2569238A1 - Electromagnet injection valve, esp. for fuel injection in IC engines - Google Patents

Abstract

The valve has an electromagnet including a core (19) and a yoke (21), an armature (23) with an auxiliary mass (22) and a valve body which is operated by the armature. The mass of the valve body is smaller than that of the armature and is not fixed to the armature, so that the power from the armature is only effective in one direction on the valve. Before the start of the operational movement of the armature, it is held at rest by an armature return spring (28), although the retentional power of the return spring arrangement is only a fraction of the saturation inductive force. The armature movement begins as soon as the magnetic power exceeds the power of the return spring, and the armature strikes the needle valve (33) at a relatively high speed at about 30 percent of the armature stroke. A part of the work carried out to open the needle valve results from the kinetic energy of the armature and the parts connected to it.

Description

 The present invention relates to an electromagnetic fuel injection valve especially for engines with internal mNstim. This injection valve is suitable for direct injection of fuel into the combustion chamber at pressures exceeding 1,000 bar. But it can also be used in an advantageous manner, to inject the fuel at low pressure into the suction pipe.

A special electronic circuit is provided for controlling the injection valve. An appropriate control system is also proposed for the pump supplying the valve.

 In diesel engines, it is sought to obtain very high injection pressures, exceeding 1000 bar, to improve the use of fuel and reduce the size of pollutants. In general, it imposes for this purpose an injection curve which starts with a very steep slope, and which defines a strictly defined injection phase. The precise moment of the beginning of the injection and the duration thereof must be adapted to the conditions imposed for the characteristics and the performances of the engine considered.

 It is easy to vary these with an electronic control system.

 For injection at high pressure, injection systems are now used which are almost exclusively mechanical, and comprise a pump, a connecting pipe and an injector. At the beginning of the injection process, the fuel is compressed in the pump, and the energy of the pump is transmitted in the form of a pressure wave to the injector.

 During the injection operation, and following it, intense pressure waves flow by reflection between the pump and the nozzle. The amplitude of these waves can reach several hundred bars. Under the effect of these pressure waves, especially after the nozzle has closed, it may happen locally that the instantaneous value of the pressure falls to zero in the injection circuit, ie below the vapor pressure of the fuel. This causes cavitation phenomena in some parts of the injection system, with sudden and intense stresses, due to shocks that occur then.

 To obtain a rapid pressure drop at the end of injection, the injection pump generally comprises expansion pistons, associated with the high-pressure valves, to allow an increase in the volume of fuel in the piping. But it is not always possible to attenuate the amplitude of the pressure wave transmitted by reflection at the moment of the sudden closure of the injection needle, and this reflected pressure wave often causes a new opening of the needle. It is then that a second fuel spraying, particularly undesirable, occurs with a delay related to the transit time of the pressure wave; and the fuel in question, insufficiently atomized, participates in combustion in an incomplete manner.

 In injection pumps, the pumping cycle is carried out with an angular setting fixed with respect to the rotation angle of the pump. As a result, the pump suffers considerable mechanical fatigue, similar to an impact, since the increase in the internal pressure of the pump then occurs completely in a very short period of time, corresponding to a low amplitude rotation angle. But the time of travel of this small angle of rotation becomes shorter and shorter as the speed of the motor increases, while the section of the flow passages of 1? injector remains the same, and the injection pressure should in principle increase as the square of the engine speed.But this sudden increase in pressure is fortunately absorbed for the most part by the elasticity of the injection piping, and by the compressibility of the fuel.

It is none the less true that this increase in injection pressure, which depends on the speed of the engine, introduces difficult problems for the regulation of fuel injection, with a view to its proper use. . At low speed, the pump pressure is usually too low to fully lift the injection needle. In addition, the increase of effective pressure at the beginning of injection is also affected by the effect of capacity of the pressure chamber of the nozzle, where there is a certain amount of fuel.As soon as the needle is partially opened, the most of the fuel pressure opposite the seat of the injector is transformed into a speed with an internal vortex effect in the injector I1 remains only a small pressure margin convertible speed, at the exit of the holes of the injector, and the atomization thus obtained is very mediocre. These difficulties can be reduced with central-core injectors. A secondary pulverization effect is also produced by the inevitable rebound of the needle when it falls back into its seat.
With the high variat ions of the useful injection pressure, which are related to the speed of the engine, it is difficult to adapt the jet to the needs of the engine, and optimum injection conditions can thus be obtained only in a narrow range of speed and engine power.

 On the other hand, the transmission of the energy of the pump by pressure waves involves certain reaction delays at the end of the pipes, and this further complicates a good adaptation of the exact moment of the injection, according to the instantaneous needs. the engine at the operating speed considered. On a large, high-powered engine, the problem becomes unmanageable due to the length of the lines, so complicated injector systems are used, or each injector is directly associated with a separate pump. , to replace a unitary assembly, mounted on the cylinder head of each cylinder of the engine.

To obtain a better adaptation of the known injection systems with mechanical control, in order to better meet the needs of an engine, an indirect effect electronic device has been proposed for controlling the injection dose and the instantaneous precise description of the operation (DE OS 3,024,424
Al). Each injector then comprises an induction detector member, to determine the start and the duration of the injection. The signals of the induction sensors are sent together with other operating parameters of the motor, to an electronic control unit which actuates an electromagnet to regulate the operation of a mechanical injection pump of the usual type. But this system does not allow to properly control the injection operations, which are not satisfactory over extended operating ranges of the engine in question.

 To avoid the problems related to the transit of the fuel pressure waves, which are at the root of most of the difficulties of the known injection systems with mechanical control, it is possible to use injection valves whose needle is controlled directly by a device electromagnetic. In this case, the pressure chamber of the injection valve receives the fuel at a constant pressure, and therefore the pressure fluctuations are much smaller when the valve is operating; the stroke of the needle is therefore little influenced. But the realization of an electromagnet for such an injection valve presents enormous difficulties, because it is necessary that this electromagnet is fast enough, and that act effectively despite the significant hydraulic forces experienced by the injection needle; it is also necessary that the energy consumption is an acceptable cost.

 Because of the difficulties associated with direct control of the injection needle by an electromagnet, two-stage control systems have been proposed (DE OS 2 914 966, DE OS 2 927 440). The injector is then provided with a complementary injection piston, mounted directly opposite the nozzle, and hydraulically controlled, at a relatively low pressure, of the order of 100 to 300 bar. Due to this mode of hydraulic transmission, the flow in the incoming piping is increased, depending on the transmission rate. Fuel is withdrawn intermittently from the incoming piping. This results in an intermittent flow which causes pressure oscillations, with an amplitude which depends almost entirely on the speed of arrival of the fuel. In relation to a direct-controlled injector, and for the same section of the supply piping, the amplitude of the pressure oscillations is thus increased in a proportion which corresponds to the transmission ratio. And at the same time, because of the pressure which is then lower, the relative amplitude increases in the same proportions. Therefore, for the same section of piping and with a transmission ratio having a normal value of the order of 5, the amplitude of the pressure oscillations is multiplied by 25. By providing a damping capacity arranged directly in the injector only these pressure oscillations can be mitigated. But the main disadvantage of this system, compared to the case of direct drive injectors, is to be much more expensive to achieve.

 German patent application DE OS 2 949 393 describes an injection valve whose needle is controlled directly by a helicoidal electromagnet having several coils excited at the same time. To reduce the bouncing effects, the electromagnet comprises two telescopic elements, of conical profile, for frictionally absorbing the kinetic energy at the end of the closing stroke.

 Thanks to the special shape thus provided, we obtain a thin-walled magnetic circuit with low eddy current loss, with a light moving armature that allows fast movement for a large magnetic force. In addition, thanks to the elongated form of the armature, relatively little subjected to lateral forces, the moving equipment has good reliability.

 To obtain with this electromagnet fairly fast movements of the injection needle that it controls, it is necessary that the majority of the magnetic induction energy arrives at the desired moment, in a very short time. It is therefore necessary to have for this purpose, and in a very short period of time, a huge electrical energy. In steady state, the consumption of electrical energy is higher than in the case where the magnetic circuit comprises a single coil, in proportion to the number of coils. In fact, the electric excitation required for a given induction depends mainly on the length of the gap and not on the surface of the working space.

 All in all, this magnetic circuit is expensive to manufacture, core winding is complicated, and multiple air gaps require very narrow machining tolerances. In addition, the wear of the damping cones seems critical.

In order to simplify the manufacture of the electromagnet and to improve the efficiency of the energy conversion, it is advantageous to use a single-coil electromagnet provided that sufficient maneuvering force is obtained and sufficient attenuation of magnetic and eddy current losses. Because of their simplicity, the forms of revolution are advantageous. The known electromagnetic injection valves which will now be described comprise a single winding electromagnet.
These known electromagnetic injection valves all have a solid magnetic circuit closed, low hysteresis and high permeability, with at least one working gap that operates at the opening; it is in this magnetic circuit that most of the mechanical stress responsible for the movement of the armature is created. In the rest of this presentation, these gaps are called work gaps. To prevent sticking of the armature in the high position, under the effect of the residual magnetism, the magnetic circuit is made so as to leave a small gap in the upper position of the armature. For this purpose, the stroke of the armature is mechanically limited, or a radial interstice is provided around it. In the case that follows these gaps are called residual air gaps. The same result can be obtained by applying to the armature and to the poles of the magnetic circuit a layer of thin coating, of a non-magnetizable substance, which at the same time improves the resistance to wear and corrosion.

 It is known that between smooth surfaces hydraulic adhesion forces occur. To mitigate this effect of hydraulic adhesion and improve the wear resistance of the surfaces in question, it is recommended (DE OS 3,013,694) to ensure a roughness of the order of 0.5 micrometer at the location of contiguous surfaces of the core or armature. One of the two joined surfaces should be as smooth as possible. The complete explanation of the physical phenomena involved in this gap is beyond the scope of our exposition.

It is generally thought that the cross section of the pole pieces should always be small or at least not increased in the vicinity of the working gaps. Thus always obtained 1 right of each gap, in the upper position of the armature, an induction that corresponds to the saturation regime of the magnetic circuit material. As the mechanical force increases according to the square of the induction in the air gap, we obtain the maximum possible value of the magnetic force for the saturation of the polar pieces, with a given value of the section of these pieces,
To ensure a very precise dosage of the amount of fuel injected at each operation of the valve, it is necessary that the movement of the armature has a low bounce rate. In a previous patent application, the
Applicant has already stated that it is possible to reduce the bouncing in a considerable way, by placing an additional mass between the reinforcement and its return spring, by adapting the movement of the reinforcement and that of the reinforcement to one another. the additional mass, by a judicious choice of the values of the forces and the masses in question, to ensure a dissipation of the kinetic energy of the armature at the end of the first cycle of rebounding, thanks to the opposite directions of the displacements of the reinforcement and additional mass at the time considered (German Patent Application DE-A 3 314 899.6).

It was also stated in the aforementioned application that by combining this additional mass system with a return spring whose curve exhibits a sudden variation, a very low rate of bounce is obtained with a very short return time. Some thought, however, that the realization of an electromagnetic injection valve equipped with such a combined system presented considerable technical difficulties, because of the extremely low travel of the armature, and that it was therefore preferable to adopt a spring with linear characteristic with very steep slope,
The conversion efficiency of the electrical energy is often greatly reduced by the losses due to the lines of force which pass outside the working gap, and by the losses related to the eddy currents.

Eddy current losses can be greatly reduced by using thin-walled magnetic circuits. As for the losses due to useless lines of force, they can especially be reduced by a geometrical configuration of the air gaps.

 In British Patent GB 1,459,598, and in European Patent Application EP 0 054 107 there is described an electromagnetic injection valve having a thin-walled magnetic circuit and a movable frame of flat shape. Electro-magnetic injection valves with flat reinforcement have an unsatisfactory and rather unfaithful movement because of the poor guiding of this reinforcement. And the conversion efficiency of the electrical energy is poor in such a valve, because in the closed position the magnetic circuits have significant magnetic losses, because of their double working gap, and the unfavorable location of these air gaps, below the winding.

 As the Applicant has already stated in another patent application, it is possible to significantly improve the guiding of the armature and the electromagnetic efficiency, compared with the case of the flat-armored magnetic circuits, by adopting a thin-walled magnetic circuit. having a bowl frame (German Patent Application P 3 314 900.3).

In the case of an electromagnet of relatively large size, it ensures the guiding of the armature by suspending it to a thin-walled guide tube. Losses due to unnecessary force lines are then greatly reduced compared to the case of a flat-armature magnet, but the rate of magnetic losses is still considerable, especially with small electromagnets whose frame has a relatively important race.

 Many specialists believe that in an electromagnet whose magnetic circuit has a double working gap, the lifting speed is considerably reduced, compared to the case of a magnetic circuit with a single working gap. This is the case for example of a electromagnet with a flat reinforcement, if one compares it to a plunger electromagnet. Indeed, for the same total value of the section of the pole pieces, and therefore for the same maximum effort, the single air gap has a double length of the other, while the surface of the pole pieces is divided by two. Therefore, for the same winding, the induction of the magnetic circuit is divided by four, and the rate of increase of the excitation current is fourfold. This view was still shared by the Applicant when he filed his application for German Patent DE-A 3,314,900.3, and is also for example in the German patent application DE-OS 3 149 916.

 For the reinforcement, a particularly low mass is obtained in an injection valve having a spherical reinforcement. In general this spherical armature is mounted below the winding of the electromagnet (European Patent Applications EP 0 051 009, 007 724, 0 063 052). These injection valves have a high rate of magnetic losses.

They have pole pieces of flat or conical shape. In these known injection valves with conical poles, the core is fixed on the opposite side to the pole, which leads to problems of centering (German patent application DE-OS 3 111 327).
As proposed by the Applicant, the magnetic losses are substantially reduced by arranging the working gap within the winding (German Patent Application 3,320,610.4). In the present invention, it is furthermore proposed to produce a part of the magnetic circuit in thin sheet metal elements, and to mount it in a boltier made of non-magnetic material, in order to reduce the eddy current losses.

 In the injection valve proposed in the European patent application 007 724, the spherical armature is returned to the closed position by hydraulic forces which makes a return spring unnecessary. This injection valve has a central bore with slots arranged radially. At the end of the opening stroke, the inlet to the injection nozzle is partially closed off by the spherical reinforcement, and the constriction thus formed between the arrival passage and the rest of the zone which surrounds the spherical reinforcement produces a pressure difference that creates the hydraulic forces. The magnetic circuit of this valve comprises a large residual gap, arranged in the radial direction, in which the spherical armature is centered under the action of the hydraulic forces. But the hydrodynamic oscillations prevent the establishment of a stationary flow long enough after the beginning of the opening movement, and the least eccentricity is enough to cause unfavorable magnetic forces, oriented in the radial direction. This is why one can doubt the stability and the fidelity of the movement of the armature of this electromagnetic valve.

 In order to improve the atomization effect of the fuel, in the case of injection into the suction pipe of an internal combustion engine, a secondary air stream is often provided around the jet of sprayed fuel coming out. This secondary air stream comes from a bypass downstream of the combustion engine intake air filter. The injection valve is then mounted downstream of the throttle valve of the intake air, so that the pressure difference at the throttle position is available to produce the air flow. secondary. In these known injection valves, the secondary air stream has substantially the same temperature as the secondary air stream sucked by the engine.

 These known injection valves have poor electromagnetic efficiency. However, in order to obtain sufficiently rapid maneuvers with low energy consumption, special electronic circuits are generally used to control these valves, by exciting their electromagnet with an energetic jolt of current to ensure the lifting of the injection needle, and then reducing the effective current of the winding, for example by modulated cuts, to the much lower intensity level which is sufficient to temporarily maintain the armature in the open position. The excitation is provided directly to the voltage of the on-board circuit of the vehicle under consideration.

To meet particular requirements of dynamic order, it can achieve prior excitation before the actual operating cycle. These control circuits are complicated and help to burden the cost.

It is also known from prior patents circuits intended to actuate electromagnets ensuring fast excitation, by discharging a capacitor.

So far, such a circuit has not been used to control an electromagnetic injection valve.

Particularities of the injection valve according to the invention.

In the cyclic maneuver of a current electromagnetic injection valve, four main phases can be considered
During the first phase, which immediately follows the application of the excitation current, the moving armature does not move. In the following description, this first phase is called the opening delay. The movement of the armature starts as soon as the magnetic force overrides the mechanical counterforce. The time elapsed between the beginning of the movement of the armature and the arrival of it at the end of the opening stroke is called the duration of opening.

In common injection valves, the armature is rigidly connected to the injection needle. This one thus carries out the same movements as the frame. After the excitation current has been cut off, the return movement of the armature starts with a certain delay, because of the eddy currents and the electrical damping effect of the winding. This second delay is called delay to closure. The return movement of the armature starts as soon as the mechanical return force overrides the opposing magnetic forces. The time elapsed until the return of the armature to its rest position is called closing time.

 In an electromagnetic injection valve, at least at the beginning of the excitation, it is possible to neglect the effect of the resistance of the winding on the shape of the curve of establishment of the magnetic force. This is slow to establish, and at the beginning of the movement of the armature, in the direction of the opening, it therefore undergoes a rather low acceleration, because the margin of motor effort is still thin.

Thus, according to the tensioning effort of the return spring, the beginning of the armature stroke is quite slow, although it is already a function of the cube or the fourth power of time, approximately. It therefore takes in general up to 75 70 of the opening time for the armature to have completed the first third of its race.

 In the high-pressure injection valves, at the beginning of the lifting of the needle, a very great effort is required to overcome the hydrostatic force which applies the needle to its seat. But the support force decreases very quickly, from the beginning of the movement of the needle, because there is a pressure equalization opposite the carrying face of the needle, from the beginning of the lifting thereof > and the hydraulic force is greatly reduced This is why the motor force needed to lift the needle decreases rapidly when the stroke increases, until it represents only about 10 to 20% of opening force.

In order to control the needle of high-pressure electromagnetic injection valves of a current kind, extremely energetic electromagnets are required, capable of providing a maximum force far greater than the needle-opening force required by its return spring, in order to be able to maneuver the needle at a fairly fast pace. At the end of the opening stroke, the motor force of the electromagnet goes far beyond the mechanical counterforce; it would thus be enough of a small part of the work of the electromagnet to overcome the antagonistic effort. But because of this excess margin - very important - of the motor force of the electromagnet, the return time is long. Moreover, even with prior excitation of the winding, it was able to provide the latter with the major part of
the electrical energy during the lifting phase which is short. Finally, with the edge voltage of a vehicle,
12 V in general, it is necessary to provide peaks of intensity which easily exceed 100 A.

In contrast, in the injection valve according to the invention, the kinetic energy of the armature is used to overcome the large hydraulic forces. For this purpose, it is arranged for the armature to hit the injection needle at a relatively high speed, after a displacement that corresponds to about 30% of the stroke.
the armature. This solution offers a number of advanced
It is no longer necessary for the maximum magnetic force to exceed the value of the maximum hydraulic force which maintains the needle applied to its seat. We can therefore use very small electromagnets whose frame has a mass of low value. Second, the duration of the movement of the injection needle is much shorter than the duration of the movement of the armature in the same meaning. It is therefore sufficient to apply to the winding a relatively slow excitation effect to obtain a sufficiently fast opening maneuver of the injection needle. The electromagnet is thus made to work almost completely using the possibilities of the electromagnet. The opening movement of the injection needle starts with a large initial velocity, whereby the pressure transformation takes place almost without delay in the nozzle holes. This gives an excellent quality of atomization immediately after the start of the injection. The duration of opening operation thus obtained has a typical value of about 0.2 ms, which seems to have never been equal until present with the electromagnetic injection valves of known types. Despite this very short opening time, the movements are smooth and very faithful, with a low speed of arrival at the end of the opening stroke. This improves the conditions of fatigue and wear of the parts in question.

 In order to obtain low value return periods, an additional mass spring system with a sudden variation of elastic reaction, as explained by the Applicant in its application, must be actuated in the opening stroke of the injection needle. prior German patent application 3,314,900.3. This document provides an additional mass mounted between the armature and the return spring, with an arrangement such that the additional mass separates from the armature when it has struck the needle, thus releasing the armature of the armature. return spring force.

Thus, when the frame tends to bounce, it undergoes a largely preponderant magnetic effort that dampens its rebound. The values involved in this dynamic system are chosen so that the shock which then occurs between the armature and the additional mass takes place with opposite facing speeds, so that the kinetic energy of the armature thus found absorbed to a large extent. However, there was still the belief that a very stiff reaction spring could provide a much more accurate motion than a spring with a steep reaction curve. But since then, it has been possible to show that with a system comprising a remnant residue booster sudden change, one can achieve mechanisms whose movements are very faithful, and tréspeu susceptible to minor manufacturing imperfections, or the effect of possible wear. For example, a variation of about 10% in the area of action of the main spring will cause the closing time to vary by only about 2%, to take an interesting case in practice. For this reason, it is preferable to use springs with sudden variation, which are easier to manufacture, rather than steep linear reaction springs.

 In the proposed system, the return force of the main spring is high, barely lower than the maximum magnetic force, so that the closing maneuver of the injection needle can then be carried out almost without delay, and with a strong initial acceleration. When the injection needle has fallen back on its seat, the armature separates from the needle to continue its course virtually without having slowed down, so that there is considerable excess effort to dampen the rebound. of the needle.

In addition, with a stepless return spring, the fidelity of the injection process can be improved over a series of successive operations. In this regard, the disturbing effect of a fluctuation of the voltage applied to the control electromagnet should first be studied.
In the elementary dose of fuel injected at each operation, depending on the duration of an electrical control signal, it can be considered that there are two parts, corresponding respectively to the quantity injected during the steady-state flow phases. transient and another amount injected in stabilized flow regime, called "stationary portion" .It generally regulates 12 importance of this stationary part by varying the stroke of the injection needle, to change the flow of flow through the injection valve. As for the "non-stationary" part of the fuel dose injected at each maneuver of the valve, it depends to a large extent on the dynamic behavior of the injection valve. This behavior can be modified by varying the elastic force of the return spring. In a common injection valve, which has a single return spring, a change in the spring force of this spring affects both the opening movement and the closing movement. If the spring force of the return spring is increased, the total duration of the opening maneuver increases and the total duration of the closing maneuver decreases. Since these two effects have opposite results as to the value of the dose of fuel injected at each maneuver, it is found that a large dispersion of the biasing force of the return spring is acceptable on various injection valves of a system. But because of this strong dispersion of the setting effort of the various return springs considered, the elementary quantities delivered at each maneuver by the corresponding injection valves will be identical to each other only for a certain value of the excitation voltage, corresponding to the setting voltage originally adopted for these valves.If the excitation voltage varies, there is therefore a dispersion between the elementary quantities delivered by these injection valves, which is undesirable of course,
Much more favorable results are obtained with the injection valves, each having a return spring with a sudden reaction variation, as proposed by the Applicant. In this case, it suffices to perform a spring force adjustment of the return spring. compressed in the upper position of the armature, which has virtually no effect on the effort of the relaxed return spring, in the position corresponding to the beginning of the opening maneuver. But for the pace of this opening maneuver, the effort of the return spring relaxed in the low position of the armature has an absolutely preponderant role.Thus, the aforementioned setting has a significant influence on the return maneuver, if although, even if the excitation voltage is varied, uniform variations in the injection dose are obtained on all the valves in question, and the reactions of the electronic control circuit can be adjusted accordingly.

 Until now, it had generally been thought that a rigid, indeformable abutment effect was needed to reduce bouncing. But it has been found that it is possible to improve rebound damping with soft stops. However, the stop must be carefully studied to have a natural frequency located in the vibratory zone where the rebound of the injection needle and the displacement of the associated abutment are performed in opposite directions. Otherwise, the rebounds will increase. In addition, with a flexible stop the impact effect and therefore the wear are greatly reduced.

 Under the proposed dynamic operating conditions, the maximum hydraulic force is greater than the maximum magnetic force, and the injection needle must be subjected to fuel pressure everywhere.

Of course, it would be possible to separate the top and the bottom of the injection needle by a closely adjusted guide, thereby compensating for the residual hydrostatic force in the open position of the needle by a helical spring. But with this system, one would have highly variable lifting forces depending on the fuel pressure variations, and therefore a very poor fidelity of the opening movements. In addition, the injection needle seal requires a very high precision adjustment, and in case of eccentricity of the helical spring force, the injection needle would undergo considerable disruptive forces.

If the moving parts are exposed to the full value of the fuel pressure, the various surfaces involved must be carried out in a special way, in particular in a high-pressure injection valve. When two smooth surfaces are applied to each other, the film of fuel that exists in the corresponding gap tends to move, to the point of allowing an intimate contact between the parts considered, especially if the pressure of support is high. This phenomenon is called
hereinafter hydraulic bonding. So if the moving parts of the injection valve that are intended to abut one another have smooth surfaces at this point, there is a risk of having between these surfaces an energetic bonding from the first maneuver, to the point to prevent the following maneuvers.

A careful study of the hydraulic phenomena involved in the contact areas between moving parts shows that the flow that occurs in such a place can be divided into several phases.
During the first phase, which corresponds to the still incomplete closure of the gap, the flow is almost exclusively subject to acceleration efforts.

Compared to other types of effort, the importance of mechanical response efforts is small and negligible. As the gap is closed, the flow is the seat of a loss of energy of increasing importance, due to the kinetic energy of the liquid which is repelled. This kinetic energy, almost entirely absorbed by the effect of turbulence, provides a perceptible damping of the movement of the parts bearing on one another. The mechanical reaction force increases as the square of the approach velocity of the bearing surfaces, and also as the inverse of the square of the corresponding interspecies. For annular interstices, the reaction stress varies as the cube from the reciprocal of the pitch dimension, and for round surfaces the variation is proportional to the fourth power of the diameter.

 When the gap becomes very narrow, the friction forces eventually become predominant in the flow They increase in linear function of the speed of approximation, and as the inverse of the cube of the value of the gap. Towards the end of the movement of approach, the friction resistance in the liquid is very important because of the narrowness of the passage, and the evacuation of the liquid is strongly thwarted. If at this time the approach speed of the two surfaces has not been seriously slowed down by the previous damping effect, there is an extremely large increase in negative pressure in the liquid between the two bearing faces.

And this increase in pressure combines with the compressibility of the liquid, to cause a reversal of movement, almost without losses. The Applicant employs the term "liquid elastic phase" for this phenomenon. "During this phase, pressures up to several thousand bar may occur even in a low pressure injection valve.
After this reversal of movement, the volume of the gap increases. In the case of an interstice between two parallel and smooth faces, the arrival of the liquid coming from the outside is insufficient to follow the movement and the flow is interrupted. Due to the local decrease in pressure, the air dissolved in the fuel is released and cavitation phenomena appear.

 With a gap having a suitable profile to ensure a sufficient supply of liquid, it is possible to avoid hydraulic bonding and interruption of the flow.

In this study of the hydraulic flows involved in the interstice, the Navier-Stokes equations which are applicable in fluid dynamics give complicated nonlinear differential equations, which can only be processed by numerical methods. Cgest why it is possible to state precise rules of dimensions only for a given case.

 In general, favorable hydraulic conditions are obtained when one of the two abutting abutting surfaces is milled from the inside outwards in the direction of flow, with a roughness of About 1 to 5 micrometers, while the other surface has undergone appropriate treatment, for example by polishing, to be smooth. The bearing area of the milled surface should not exceed 10. To reduce wear, the two bearing surfaces are hardened, preferably by nitriding. The abrasion grooves oriented in the direction of the flow allow the evacuation of the tiny fragments which can be detached, so as not to hinder the flow of the liquid.

 Another possible solution to prevent a hydraulic bonding is to give one of the support faces a loose shape, flange, bowl or membrane, with a slight elasticity, to obtain an annular bearing on the other side support when the interstice is gone. When the support forces change, the support faces associated abutment can separate from one another, their profiles facing then detaching from each other first edge and then gradually to interior in the direction of the flow, so as to favor the arrival of the liquid in the interstice. The interaction effects between the parts in question can be improved by giving a slightly arched profile shape to one of the two associated faces in abutment, If the associated bearing surfaces are curved, it is necessary to choose the natural frequency of the moving parts in question, by adapting them to each other, as already indicated, to obtain a shock effect with movements snapshots of opposite meanings.

 It is also possible to give one of the two bearing surfaces associated a chamfer profile, so that the section of the gap goes increasing from the center to the outside. The corresponding angle of obliquity must not exceed 1 ", and even have a much lower value.With interstices interesting large surfaces, one can thus obtain a strong damping effect at the end of the movement of opening, in order to suppress the bouncing that occurs then.

The other hydraulic effects associated with moving the moving parts of the injection valve are of very little importance, provided that passages of sufficient cross-section are provided for the pressure equalizations. This is due to the fact that the pressure anomalies are balanced at the speed of sound within the fuel. On the other hand, the individual parts have a very low maximum speed, of the order of 1 to 2 m / sec, so that it suffices to stick to the hydrostatic effects to study the effects of the movements of these parts on within the liquid fuel, except for the hydrodynamic phenomena involved in the interstices of the
companies in abutment.

 Of course, intense hydrodynamic oscillations can occur, but they have little influence on the movements of the various moving parts of the injection valve. These oscillations can be used to act judiciously on the injection operation. But it must be verified that these oscillations occur only at the inlet of the injection nozzle, without inducing any coupling effect in the connecting pipes mounted between the various injection valves of the engine in question. It is indeed necessary to stabilize the fuel pressure throughout the system upstream of the injection valves by avoiding compromising the fidelity of the injection operations performed by each valve. For this purpose, it is advantageous to provide compressible elements in the immediate vicinity. injection nozzle or shutter of each valve.

Since the amplitude of the pressure oscillations depends directly on the flow velocity of the fuel, the fuel supply passages to each of the injection valves must be given as large a section as possible.

 In the low pressure injection valves provided for fuel injection into the engine intake manifold, the atomization quality of the fuel can be improved by providing a stream of atomizing air, such as a known way. In the known injection valves of the kind in question, the atomizing air stream comes from a bypass taken downstream of the intake air filter of the engine. The injection valve is mounted downstream of the throttle valve of the engine intake air; this is to take advantage of the pressure difference created by the throttle to circulate the atomizing air flow. But when the throttle fully opens, the pressure difference disappears almost completely, and the circulation air flow is practically interrupted.

 On the other hand, when the throttle is open, the intake air current is intense, and causes a significant pressure drop at the intake air filter.

Thus, the suction pipe of the engine is the seat of sufficient vacuum with respect to the outside, at least during important periods of the operating cycle of the engine, and it can be used to obtain an air flow of atomization having a high speed. With a depression of 50 millibars, for example, it is possible to have a current of air of some 100 m / s, and this speed makes it possible to improve the atomization in a very satisfactory manner.

 It is also possible to use the depression of the suction pipe of the engine when the throttle is fully open by causing the atomizing air flow to arrive from a separate air filter, reserved for the atomizing air stream. This arrangement is particularly interesting because the large depressions of the suction pipe are related to the high speeds of the combustion air, and therefore to a significant throttling effect at the location of the intake air filter of the motor, while at the location of the atomizing air filter the throttling effect tends to decrease, because the speed of the atomizing air current is lower.A particularly simple solution and effective is to arrange the separate filter of the atomizing air stream directly on the injection valve, passing the atomizing air stream through the valve area where the valve is located. winding of the electromagnet, which at the same time makes it possible to improve the cooling of this winding.

 The effect of atomization of the fuel and thus the efficiency of the engine can be further significantly improved by heating the atomizing air. For this purpose, it is possible to provide a heat exchanger, constituted for example by a coil wound in a helix, which is arranged directly in the stream of hot engine exhaust gases. This heat exchanger is placed between the air filter and the atomizer device. Thus, when the throttle valve is closed, the combustion air that arrives at the engine consists almost exclusively of high temperature atomizing air. This ensures an excellent fogging effect of the fuel whose condensations on the intake manifold are reduced accordingly. The high temperature of the intake air reduces the ignition time of the engine at reduced power, which therefore improves the efficiency of the engine. By thus improving the combustion in the engine, the operating range can be extended to a lean mixture, thereby reducing pollutant emissions. When the throttle valve is then opened gradually, the high temperature atomizing air stream is mixed with an increasing proportion of cold air, so that the temperature of the combustion air decreases. This makes it possible to maintain a sufficient margin with respect to the limit of detonation of the engine. At full opening of the throttle, the heating of the combustion air is negligible, because of the low proportion of atomizing air.But we still obtain at this regime a strong improvement of the atomization effect, at because of the high temperature of the atomizing air. On the other hand, the speed of the atomizing air stream is greatly increased by the heating of the air, especially at low pressure differences, since an increase in air temperature for the same difference pressure always causes a sharp increase in the flow velocity. In addition, since the fuel mixture thus obtained is very well adapted to the needs of the engine, the range to be provided for the setting of the ignition point is much narrower. The heating of the atomizing air can, however, give rise to serious problems with regard to the injection, because of the vapor bubbles which tend to form within the fuel in the injection valve. the appearance of these vapor bubbles, it is therefore always necessary to provide a heat-insulating protection of the injection valve in connection with the atomizing device, and to ensure additional cooling of the injection valve, by circulating a fresh fuel flow.

 At high powers, the performance of internal combustion engines is limited by the appearance of detonation phenomena. On modern engines the effects of detonation are avoided by reducing the ignition advance, depending on the reactions of a detonation detector. But by reducing the ignition advance, it also reduces the efficiency of the engine. Or by injecting water into the engine can improve its performance at high power regimes because it greatly reduces the extreme combustion temperatures without sacrifice on the intrinsic efficiency of combustion in the engine. This reduction in peak temperatures allows for a large reduction in the nitrogen oxides, and in general it is possible to avoid having to reduce the ignition rate. It can even be increased. An excellent adaptation of the operating characteristics of the engine can be obtained by injecting low pressure water into the suction pipe of the engine, by means of an electromagnetic injection valve. slave to a detonation detector. With this arrangement, it reduces the water consumption, and as the water injection takes place at high power regimes, and therefore at high temperature of the engine, the water is not likely to condense in the engine and there is no risk of corrosion. There is no need to stipulate special conditions as to the quality of spray water, because it necessarily enters the engine in droplets large enough to vaporize only at the end of compression and during combustion.

It is therefore sufficient for the supply of a simple float tank associated with a nozzle, such as in a carburettor, with a solenoid valve controlled by the detonation detector.
Tests carried out by the Applicant have shown that with an injection of water, the combustion takes place practically without residues, and thus almost completely avoids deposits of combustion residues in the engine.

In addition, the use of the injection system in question on an internal combustion engine allows almost unlimited intake compressor overcharging, up to the practical limits of mechanical strength of the engine. In a supercharger engine water injection is always practiced upstream of the compressor, to improve the atomization by stirring and also improve the efficiency of the compressor.

 Even with standard-type combustion engines, the injection of water makes it possible to use fuels with a very low octane number, without sacrificing the compression ratio of the engine. An excellent adaptation of the characteristics of the engine is obtained by providing a water injection associated with the hot air atomization system described above.

 In order to have a good fidelity as to the accurate metering of the elementary quantity of fuel delivered at each maneuver of an injection valve, it is obligatory to calibrate the injection valves usually with fuel. The production of low pressure injection valves is greatly simplified if this acda $ air calibration is performed. With regard to the "stationary part" defined above for the elementary quantity of fuel discharged at each operation of the valve, identical conditions of flow for air and for fuel are obtained if the same value is used. In addition, operating with air requires a largely subsonic flow rate to have comparable conditions. it is therefore sufficient for a pressure difference of some ten mbar to achieve the flow of air. But under the ambient conditions, the kinematic viscosity of the air is much higher than that of the fuel. The kinematic viscosity of the air can be reduced by increasing the pressure. In general, an air pressure of 5 to 10 bar is sufficient, it is therefore a much higher value than a low pressure fuel injection valve, which operates between 0.7 and 3 bars approximately .

 The opening maneuver of an injection valve is generally strongly influenced by the hydrostatic effects. This is why the calibration of an injection valve, as regards the "non-stationary" part of the injection dose, directly related to the dynamic behavior of the valve, must be carried out at a pressure of air that corresponds to the fuel injection pressure. In doing so, the effects of damping the fuel on the operation of the valve and the effects of the hydrodynamic oscillations are obviously not taken into account. However, we find well the running tips of the various phases of movement, which have the most importance for the "non-stationary" part of the injection dose.With this calibration method, we can take into account the possible differences in introducing appropriate correction coefficients To measure the displacements of the armature and to study its movement one can resort to photocells, for example, or refer to the electrodynamic reactions of voltage in the leaning of the electromagnet.

 In the proposed injection valve, the kinetic energy of the armature is used to overcome the force of the opening, and this makes it possible to have short opening times for the injection needle, with relatively long durations for the movement of the frame in the same direction. This requires a small increase in the induction rate in the magnetic circuit. Under these conditions, the rate of eddy current losses is also very low, and one can therefore have a relatively thin-walled magnetic circuit. With this low loss rate, the value of the maximum intensity is divided by approximately ten. power supply, compared to a current solenoid valve.

 In theory, and for the same initial value of the inductance, the magnetic stress setting pace does not depend on the fact that the electromagnet has a simple working gap, or a double gap0 Indeed, everything depends in principle the accumulated energy that corresponds to the magnetic field and the stroke of the armature. For a given operating time, the electric power consumption therefore depends solely on the initial value of the inductance of the electromagnet, neglecting the resistance of the winding.

 With an electromagnet with a double working gap, the number of turns of the excitation winding must be multiplied by four to have the same inductance as in a magnet with a single working gap. Therefore, for the same current path and the same current density, the section of the winding must also be multiplied by four.

In addition, the section of the blades is reduced by half, and the total length of gap is doubled. All this greatly increases the reluctance of the magnetic circuit, and therefore the rate of magnetic losses of the electromagnet. However, the magnetic loss decreases with the square of the loss rate, so that the magnetic losses are of great importance for the dynamic behavior. The magnetic losses increase the inductance of the winding and strongly reduce the magnetic stress in saturation regime, in the low position of the armature.

 On the other hand, in a double working-gap electromagnet, the eddy current losses are reduced by about three quarters, since the wall thickness of the magnetic circuit is twice as small. To have a sufficiently low rate of eddy current losses, the wall of the magnetic circuit must have a thickness of 0.5 to 1 mm maximum.

 But with a thin-walled magnetic circuit, in a current-size injection valve, in the lower position of the armature, the magnetic resistance of the air gaps is much greater than that which exists between the core and the cylinder head. It is thus formed an intense field of leaks, which bypasses the gaps.In the low pressure injection valves whose electromagnet comprises a flat armature, this field of leakage may represent for example, for a magnetic circuit of current dimensions, up to 75% of the total flux, which reduces by the same efficiency of energy conversion in the electromagnet Or the dynamic behavior of the electromagnet depends mainly on the speed of establishment of the field magnetic at the beginning of attraction, and it is therefore very important to reduce the field of leaks to obtain fast movements, with a low rate of losses. To obtain favorable yields, it is necessary to use certain special configurations for the pole pieces. With small values of the polar sections, it is advantageous to use electromagnets at a single working gap, because of their reduced reluctance. it is then necessary that the working gap is located substantially in the center of the winding because there exists in this place a concentration of the lines of force allowing a conversion of energy at a low rate of losses. With electromagnets with a double working gap, the best performance is obtained with a cup-shaped frame, whose profile matches that of the winding, and whose poles are arranged so as to cover about a quarter of winding (German Patent Application 3,314,900.3).

With elongated form windings, a reduction of about 75% is obtained for the leakage field, compared to the case of a flat-armored electromagnet, with the same polar section value and a winding of same dimensions. With this arrangement of the blades, one obtains a performance as good as in an electromagnet with a simple working gap arranged in the center of the winding, but one then takes advantage of a polar section reduced by half, which thus strongly reduces the eddy current losses for the same value of the magnetic force.

 In cup-inset electromagnets, difficulties are encountered in sealing and fixing the coil. In this respect, it is advantageous to use a magnetic circuit with a double working gap, in which the outer pole of the armature is constituted by a collar of small diameter. Electromagnets of this type are described in German Patent Application OS 3,149,916 and in European Patent Application 0 076 459.

In both cases, these are electromagnets having a short armature, and whose poles are located below the winding, that is to say with a large loss field.

Particularly in the case of the injection valve described in German patent application OS 3 149 916, it appears that with a relatively thick wall magnetic circuit it is difficult to have an improvement over the injection valves known to simple working gap. The solution in question, however, has an advantage because the attraction force is established almost without transverse component, even if the frame is suspended with a slight eccentricity, which is always possible.

 With these electromagnets, much better yields are obtained by placing the inner pole above the center of the coil. We have a maximum attraction force when the polar sections have substantially the same value, and placing the inner pole substantially at the height of the upper quarter of the winding. For low-pressure injection valves of appropriate dimensions, it is often sufficient to have a low-value attraction force, which can be obtained with a single working gap and a magnetic circuit whose wall has a thickness of the order of 0.5 mm. Again, a double air gap is interesting to ensure the suspension of the armature without transverse component.For this purpose, we can increase the section of the outer pole, in order to reduce-the reluctance of the corresponding gap and thus reduce the field leaks.

With this configuration, the best performance is obtained by arranging the inner pole substantially in the center of one winding.

 With antagonistic mechanical forces of low value and a very light reinforcement, it is sufficient to have an equally weak attraction force. Thus, in a low-pressure injection valve, unlike the solutions usually adopted, it is possible to go up to enlarge the polar section relative to the current section of the magnetic circuit, in order to reduce the reluctance of the gaps and therefore the field of view. leaks, thereby reducing losses by currents of Foucauît With such a configuration, we obtain a conversion of energy with almost no losses. For the same conditions of thermal fatigue and inductance as in a conventional electromagnet, this reduced reluctance makes it possible to have much smaller windings, with a low number of turns. But on the other hand, with the usual dimensions. , the polar section is not greater than that of the rest of the magnetic circuit, in order to have a low leakage field, and this gives in the saturation regime significant attraction forces, much higher than the antagonistic and driving force therefore a significant time for the return maneuver. It is sought to reduce this duration by reducing the intensity of maintenance with an electronic device. On the contrary, with the proposed solution, it is possible to use simple actuation circuits without artifice for reducing the sustaining intensity, and the dynamics of the system is improved at the same time. In addition, the proposed solution also makes it possible to perform a simple modification, to improve the electromagnetic injection valves already in use, because it is then enough to reduce the section of the core above the pole by a simple drilling.

 Another significant reduction in the leakage field of magnetic circuits of usual dimensions can be obtained, especially on small electromagnetic injection valves, by providing large openings in the skirt of the magnetic revolution circuit, which completely surrounds the winding. In doing so, the reluctance between the skirt and the core is increased, which has the effect of reducing the field strength of leaks.

 In the high-pressure injection valves, to obtain sufficient attraction forces, it is necessary to provide important polar sections, the corresponding air gap thus having a low reluctance. Thanks to the provisions proposed above, it is possible to further reduce the leakage field, to the point of obtaining a fairly high electromagnetic efficiency, even with materials of very low permeability. It is thus possible to use composite materials comprising a low hysteresis powder which is coated with insulating plastics material. These materials have a high electrical resistance, which almost completely prevents the appearance of eddy currents. But in general their maximum relative permeability can hardly exceed a value of 200 to 300. With these materials, one can achieve compact magnetic circuits, of sufficient strength, able to withstand high pressures. To have a reliable suspension of the armature, it is fixed to an elongated thin-walled guide tube which serves at the same time abutment, not to subject this relatively soft magnetic material to excessive stresses. The reinforcement can be made by stamping together with the guide tube, to form a single piece therewith. To increase the magnetic flux, the thin-walled tube can be made from a low hysteresis material, nitrided hardened to improve its wear resistance. This hardening treatment hardly affects the low-hysteresis qualities. With windings having a sufficient compressive strength, the magnetic material can be compressed directly around the winding, which facilitates tight sealing of the material, simplifying manufacture.

 To make high-pressure injection valves, conventional coiled wire windings are poorly suited. It is sufficient to provide a small number of turns therefore a small number of levels for these windings for the small inductances that are necessary. Between the ends of these levels, induction maxima appear, especially because of the extra-breaking currents, and the insulator of the wound wire can then suffer. The windings made with a thin ribbon are much more interesting because they withstand much higher mechanical and electrical stresses. For mass production, these coiled ribbon windings are also less expensive than windings wound wire. To achieve these windings can be taken for example a thin strip of anodized aluminum surface, with which an intermediate insulating layer is superfluous. It is also possible to remove the profiled frame with such a winding, in favor of a better use of the volume of space available. Preferably, this winding is impregnated with plastic under vacuum, which improves its strength. The contacts at the ends can be provided for example by means of sockets slit lengthwise, which further improves the solidity of the set. It is also possible to fold down the ends of the conductive strip, by projecting them at right angles to the turns. With a coil which thus has sufficient strength, it is advantageous to coat it directly with the magnetic composite material based on powder. , performing several operations as needed.

 If the winding housing volume is to be sealed, it is advantageous to use a ceramic profile mount for winding. The novel high strength ceramic materials developed for the manufacture of engines and turbines are preferably used. To improve the mechanical strength of the profiled mount, it is good to wind the tape on its mount by stretching the tape to the maximum, to introduce a preload effect in the frame.

Some embodiments of the invention will now be described by way of example, with reference to the appended drawings, in which
- Figure 1 shows a high pressure electromagnetic injection valve according to the invention;
- Figures 2a, 2b, 2c are graphs for explaining the operation of the valve of Figure 1, during its lifting movement; Figure 2a showing the forces involved, while Figure 2b shows the speed variation of the armature, and Figure 2c the movement of the additional mass of the valve;
3a, 3b, 3c are detailed diagrams of a double-gap internal electromagnet provided for an advantageous embodiment of the valve according to the invention;
- Figures 4a and 4b are detailed diagrams of an improved electromagnet for the injection valve according to the invention;
- Figure 5 shows a low-pressure injection valve, for combustion engines where the injection takes place in the suction pipe;
- Figure 6 is a diagram of the forces involved in the valve of Figure 5;
FIG. 7 represents a spherical electromagnetic injection valve;
FIG. 8 represents a spherical electromagnetic injection valve comprising an atomizer device;
FIG. 9 represents an electromagnetic injection valve provided with a hot air atomizer device;
FIG. 10 is a diagram of the control circuits of an indirectly adjustable injection pump;
FIG. 11a is a diagram of a control circuit associated with two electromagnetic injection valves;
Figure 11b is a diagram of another control circuit for any number of electromagnets.

 FIG. 1 shows a high-pressure injection valve according to the invention. The operation is explained below, with reference to a characteristic cyclical movement, the lifting phase of which is studied in FIG. 2.

 The injection valve of FIG. 1 comprises an electromagnet whose magnetic circuit, made of composite powder material, is constituted by a core 19, a yoke 21 and an armature 23. The winding 18 is housed in a shaped support in ceramic 20. The core 19 extends almost to the lower end of the support 20 of the winding, to mechanically support it. It is therefore possible to use a ceramic material of relatively low strength to carry out the support 20 of the winding. The magnetic circuit has only one working gap, to obtain a polar surface as large as possible. That is why, despite the position of the work pole, located below the winding, which is In principle, despite the low magnetic permeability of the electromagnet material, leakage rates of a still acceptable value are obtained. Thanks to the small diameter of the lateral pole, the lower face of the winding support is completely covered. This makes it possible to apply to the entire magnetic circuit an energetic compression force, in the longitudinal direction, with a view to sealing them reliably. The use of a sealant or adhesive may facilitate this sealing operation.

 The electromagnet is housed in boltier 16. Preferably, it is made of austenitic cast cast high strength, non-magnetizable. The housing 16 is closed by a screwed lid 13. The stop 17 serves to determine the residual air gap that remains under the central pole when the armature is raised. The additional mass 22 is suspended from the abutment 17, the cup-shaped portion above the electromagnet is also used to protect the relatively soft powder composite material to avoid damaging this material when we close the case by screwing the cover in place. The core 19 is fixed firmly to the abutment 17, preferably by gluing, or directly by extrusion during the manufacture of the core, to be able to machine at the same time the polar surface of the core and the bearing surface of the abutment, taking the coin in a single vise. In addition, it is advantageous to shield the armature 21 at the point of its portages in the casing, by means of a firmly fixed plate, to reduce the risk of damage to the assembly To reduce the reluctance, it is advantageous to can be used plates made of a low hysteresis material, which is cured superficially, preferably by nitriding, to improve their resistance to wear.In the proposed embodiment for the magnetic circuit, it is possible to Having a small internal diameter for the high-pressure portion of the injection valve, whereby the mechanical stresses of this part are reduced. Thus, it is possible to use a compact housing with a relatively thin wall.

 The additional mass 22 projects slightly protruding over the retaining face of the abutment 17, so as to obtain a sudden variation in the curve of the elastic resetting force. The importance of this protrusion of the additional mass is chosen so that the force of the main spring 15 acts towards the end of the lifting stroke over a distance of about 30 to 50% of the total stroke of the needle of the injection valve. The importance of this projection is not absolutely critical, so that it can be avoided if the manufacture has been carried out to a suitable degree. By acting on the adjusting screw 14, it is possible to adjust the elastic force of the main spring 15, and thus also the resetting force towards the end of the lifting stroke of the needle. At its lower end, the screw of adjustment 14 carries a tubular guide piece 26 on which are mounted two springs 28 and 35, low reaction. These two springs have characteristic low slope curves, so that the elastic force of each of these springs varies little when the adjusting screw 14 is rotated. The inner spring 35 serves to apply the needle 33 of the valve. injection in its seat, even in the absence of pressure in the system, to ensure in any case a proper bearing seal of the needle 33 on its seat, even when the engine is stopped. The outer spring 28 serves to rearm the armature, providing a force at the beginning of the movement of the armature, to prevent it from bouncing against the needle at the end of rearming, which would then cause a new unwanted uplift of the armature. injection needle. The force of the resetting spring of the armature is transmitted by the cup 29 to the connecting washer 30 which is disposed in the thin-walled tubular guide 24 of the armature.

 The useful stroke of the armature and that of the injection needle is adjusted by means of washers of appropriate thickness, namely a setting washer 36 for the travel of the armature, and an adjusting washer 37 for the stroke of the needle. These two adjusting washers are firmly applied against each other, resting on the body 31 of the nozzle by means of a clamping sleeve 27. The injection valve is mounted by screwing into the cylinder head of the motor acting on the profiled head 25. Like a nut The injection needle 33 slides in the guide 32 which ensures its guidance.

The guide 32 of the injection needle may be provided with discharge notches, which provide a uniform distribution of pressure in the gap of the guide. This particular arrangement is of interest in the case of the injection valve according to the invention, unlike the common injectors, because the operating conditions are substantially different in the case considered. In addition, it is possible to mount the injection needle with a relatively large clearance, to ensure, within certain limits, a spontaneous centering of the needle. The embodiment of this guide associated with the injection needle according to FIG. The invention requires much less precise machining accuracy than for conventional mechanical injectors, since the injection guide in question is not required to provide a special sealing function.
The body 31 of the nozzle has at its bottom a relatively thin wall, in order to have at this location a low enough natural frequency. We choose the value of this natural frequency, to further reduce the bouncing of the injection needle. This rebound itself already has a duration which is of the order of a microsecond only, but it is further reduced by In addition, this flexible shape reduces the mechanical fatigue of the seat of the injection needle.

 In the proposed injection valve, by appropriately selecting the diameter and length of the seat supply channels of this valve, as well as the fuel volume below the injection needle guide, it is possible to get almost any injection process at will. In the injection valve shown, the feed channels drilled in the guide 32 of the valve needle were made with relatively thin sections. This results in a very marked pressure drop at the opening of the valve, and this results in high pressure oscillations during the injection. This mode of operation may be favorable for some engines. The oscillation frequency is then essentially determined by the length of the arrival channel. With a short channel, a relatively low frequency oscillation regime can be achieved by using the capacity resonance of the fuel volume below the guide of the injection needle.

This arrangement can be used to achieve a pre-injection effect before the main injection. But in general, we will achieve arrival channels with a caliber as strong as possible, so as to obtain an injection regime practically without oscillations, with a pressure curve with a steep upward slope, in order to reduce the mechanical forces necessary to lift the injection needle.

In the injection valve shown, there is further provided a damping member 34, plastic much more compressible than the fuel. This ensures an attenuation of the pressure oscillations and an accumulation effect. In addition, the residence time of the fuel in the injection valve is shortened. However, the use of such a damper is only of interest for relatively low fuel pressures
We will now study, with reference to Figure 2, the graph of the movement of the injection valve of
Figure 1. All characteristics are represented at the real scale of this cyclic movement.

FIG. 2a shows the variation of the magnetic force F and the sum of all the ef
Fmech strong mechanical resistance mag, depending on the value "s" of the moving armature stroke. It can be seen that the magnetic force increases very rapidly as the stroke increases. At first glance, it is surprising, since the magnetic strength increases approximately like the square of time, and thus very slowly at the beginning. But this slow increase in magnetic stress is associated with a slow increase in the acceleration of the armature, so that the stroke of the latter is low at first. However, despite the slow establishment of the magnetic force There is significant kinetic energy to overcome the resistance force when opening the valve to lift the injection needle shortly after the start of the stroke thereof.

 The movement of the armature starts as soon as the magnetic force outweighs the elastic force of the rearming spring of the armature. At the end of a race sl, the armature strikes the injection needle. The integral of the useful work for the acceleration of the armature for the lifting of the needle is represented by the hatched areas of Figure 2a.

 The graph of Figure 2b shows the variation of the speed of the armature, depending on the stroke "s" thereof, and Figure 2c shows the variation of the stroke "s" of the armature, depending time "t". It can be seen that even at the point of its short sl course, the speed of the armature has already reached more than half of its final value, but for this brief race, it takes a delay t1, which is very long, because it represents much more than half the total duration of the uprising movement.

 When the armature strikes the injection needle, this impact causes a loss of speed avl which is very low, because of the great difference which exists between the mass of the armature and that of the needle. mechanical counterforce increases sharply and greatly exceeds magnetic stress. The work that results from the kinetic energy of the moving parts is represented by the cross-hatched area of Figure 2a. This results in a slight reduction in the speed of rotation. As soon as the mechanical antagonistic force has fallen below the magnetic force, the speed starts to increase again.

When the armature has traveled the stroke s2, the moving parts strike the additional mass at time t2, which causes another impact effect with a slight loss. The speed decreases slightly, to increase again with a lower gradient. The opening movement of the injection needle ends at time t3, which represents a relatively short duration tA for this opening movement
At time t3, the armature hits its stop and bounces. This causes a considerable loss of energy and speed because the stop is fixed and therefore immobile. But the additional mass continues freely, and thus relieves the armature of most of the effort of the return spring. Thanks to this, the bouncing phenomenon of the frame is strongly abbreviated, and if we chose correctly the value of the additional mass, the residual kinetic energy is dissipated to a large extent in a new shock between the reinforcement and the additional mass animated with opposite velocities. The path of the additional mass is shown in Figure 2c by a dashed line. Figure 2b shows the magnitude of the loss of speed. For all these reasons, an extremely fast and gentle movement is obtained, and the mechanical fatigue imposed on the structural parts is much lower, thanks to the low value of the maximum speed, compared to a current injection valve.

 In the case of the injection valve of FIG. 1, an electromagnet of a rather unfavorable yield has been used, but with which it is possible to have a winding mounted on a ceramic profiled support of rather low strength. In the diagrams of Figure 3, a few more favorable design details for such an electromagnet have been indicated.

 FIG. 3a shows an electromagnet with a double working gap. This electromagnet comprises a core 40, a winding 41 and an armature 43. The outer working gap is made with an oblique profile, in order to obtain a low reluctance associated with reduced transverse forces, in the event of eccentricity. . To further reduce the radial forces and to have a frame of smaller footprint can be provided an outer working gap having a stepped profile, with at least two steps. The magnetic circuit consists of powder-based composite material.

 Figure 3b shows an electromagnet for high-pressure injection valve, made of powder composite material, with an outer pole flange. This electromagnet comprises an armature 52, a tubular guide 53 and a yoke 46. The winding 50 consists of a ribbon connected to two split metal rings 48 and 49.

The winding mount 47 is made of high strength ceramic. The inner pole is located at the most favorable place for the construction of the electromagnet, so that the frame covers about 3/4 of the length of the coil. The cylinder head 46 is reinforced by a metal washer 51 at its base, and by the stop 44 at the top. The stop 44 serves at the same time to support the additional mass 45.

The stop and the coil are mounted together in a single operation, and fixed to the cylinder head, the tubular guide being likewise fixed to the frame. On the periphery, the cylinder head has large openings, to reduce the field of losses.

 Figure 3c is a detail enlargement of the abutment faces of the electromagnet of Figure 3b. The tubular guide 53 is provided with radial grooves ensuring a pressure equalization at the closing of the corresponding gap. The surface of the guide tube is hardened and ground. The contact faces of the abutment 44 and of the additional mass 45 are chamfered on both sides, to prevent a hydraulic gluing effect. Compared with a chamfer made on one side, the chamfers on both sides reduce mechanical fatigue. abutting contact faces. Of course, the contact faces can also be rectified in the radial direction, as already ex-posed.

 Figure 4a shows an electromagnet with external pole flange, material low hysteresis. To simplify manufacture, the armature and the guide tube form a single piece 66. As the permeability is not too high for the material of the magnetic circuit, the armature is preferably made of annealed special steel2 of high electrical resistivity. hardened for example by nitriding or pitch coating. in amelioriocertla-resi2nce sqerficielle to wear. To reduce eddy current losses, the armature can be split lengthwise. The wall of the armature preferably has a thickness of between 0.5 and 1 mm. In order to obtain greater forces and thus to have a thicker wall, it is possible to make the armature re by assembling two or more isolated sleeves and The return of the magnetic flux passes through the core 60, the perforated skirt of large section constituted by the two concentric rings 61 and 62, and the lower washer 65 of the yoke. The coil 63 constituted by a ribbon is reinforced by a tubular ceramic mount 64.

 To further reduce the losses due to eddy currents, the electromagnet of Figure 4b is formed in part of laminated pieces. The armature 72 made of composted powder-coated material on the guide tube 701, which can also be made of low hysteresis material, the inner pole is located above the winding, to facilitate the manufacture of this piece 67 Even for electro-magnets with a strong electro-magnetic effect, usually only 2 to 4 strips are required. However, there are intense eddy currents in the electromagnet considered, especially in where the strips bear one on the other, because the orientation -lattice-does not coincide; then with the direction-of the flow my magnetic. The corresponding losses can be reduced by arching or bordering the strips in the flow direction. or by practicing flip flop on these bottles, but it is to the detriment of the cost price.

Figure 5 shows a low-pressure injection valve for combustion engines where injection takes place in the suction pipe. The magnetic circuit consists to a large extent of thin strip elements, having walls of about 0.5 mm thickness. The core 83 is crimped on a tubular appendage of the cover 80, non-magnetizable material. This gives better mechanical stability and a satisfactory centering of the core. The yoke 86 is provided with openings of large area, to reduce the field of losses. The cross-section of the electromagnet is substantially square, and this form, which is the most favorable for making a high tro-magnet, ensures a reduction. additional loss field The reinforcement and the guide tube form a single piece 88. To help further reduce the currents of
Foucault and ensure pressure balancing, the frame is slotted in the direction of the length. The outer pole has a much larger cross section than the inner pole, to reduce the reluctance of the magnetic circuit. With this arrangement, it is sufficient to provide for the winding 85 a relatively low number of turns in order to obtain a positive effect. induction sufficiently high in the electromagnet. Thus, for a given cross-section of the winding, the thermal fatigue thereof is reduced.

The outer side, the yoke and the core are superimposed and are held firmly compressed by the cover 80, bearing against the bottom of the housing 87, also consisting of a non-magnetizable material.

 The injection valve is provided with a shutter member 94 in the form of a cap, and of relatively large diameter. This large diameter makes it possible to have a favorable shape for a good flow, with a seat of equally large diameter, by virtue of which it suffices to give this shutter a fairly low stroke, even for a high flow of fuel. The shutter 94 of the valve is mounted in the guide tube of the armature with a reduced radial clearance, to ensure self-centering. The shutter 94 is provided with a number of radial holes of large diameter, to ensure the flow of the fuel with a. throttling reduced.

 The flange of the frame rests on the body 92 of the injector, with a large bearing surface, to ensure a hydraulic damping of the movement of the armature, at the return thereof. To avoid a hydraulic gluing effect, the nozzle body is milled in the radial direction. Pressure equalization is also provided by radial holes in the guide tube. When the armature is in the low position, there is a slight axial clearance between the armature and the shutter 94, to allow armature to start its race. The armature is biased down by the spring 82, while the shutter 94 is recalled by the spring 91 which is much more energetic. The spring 91 acts on the upper part of the shutter, so as not to hinder the flow of fuel. But this assembly can cause undesirable radial forces, in case of relative eccentricity. These undesirable radial forces can be reduced by terminating the spring within the obturator 94.

 Adjustment of the initial stroke of the armature is done by matching different reinforcement or shutters. The opening stroke, and thus the value of the full-opening fuel flow, is adjusted by pressing the shutter 94 more or less deeply into the casing 87. The final force of the return spring, and therefore the part of the amount, is adjusted. of fuel injected at variable speed, by modifying the position of the adjustment tube 81. The spring 91 has a yield curve of very high upward slope, while the spring 82 has a low slope curve, so that the adjustment of the elastic return force stobtient almost solely by acting on the spring 91, the initial elastic force varying little at the beginning of the course of the armament.

 The injection valve according to the invention has a very important section for the arrival of the fuel passage, and the speed-flow of fuel is low. Thanks to this low speed of arrival of the fuel, hydrodynamic pressure oscillations are greatly reduced, during the operation of the valve, with respect to the case of known injection valves, where the flow of fuel occurs. At the same time, the oscillations are eliminated almost completely, thanks to a damping volume provided around the body of the injector, in the immediate vicinity of the shutter. This damping is ensured by the elasticity of the sheath 93 mounted around the damping volume, and which serves at the same time as a seal between the housing and the body of the injector, and thermal protection of the valve in the suction pipe of the engine. Steam bubbles that may form may escape upward, passing through axial grooves in the body of the injector. As for the vapor bubbles that gather at the top of the injection valve, they are discharged through radial holes of the control tube 81, under the effect of the depression due to the flow of fuel.

 It is also possible to damp the hydrodynamic oscillations by associating a rigid wall with the damping chamber which is then produced in the manner of a resonant cavity, also called "Helmholtz resonator". A resonant cavity is an enclosure provided with at least one opening, and which thus has a characteristic natural frequency which depends on its dimensional disposition.

The natural frequency of the resonant cavity considered is adjusted to match that of the strongest oscillation occurring in the injection valve during operation to eliminate this oscillation to a large extent. The only requirement for the efficiency of this resonant cavity is that all the dimensions of this cavity must be less than one quarter of the wavelength of the corresponding resonant frequency. For the evacuation of the steam bubbles, it is also necessary at the top of the valve purge holes, or purge grooves, as indicated above, but it is necessary that the cross section of these evacuation passages is very low , so as not to reduce the effectiveness of the resonant cavity. The dimensions of realization to be provided for this resonant cavity are indicated in the works and technical publications where this question is treated.

 The graph of Figure 6 shows the variation of the magnetic stress and the mechanical counterforce of the injection valve of Figure 5, as a function of the lift stroke "s". "S". The movement of the armature begins at the moment when the magnetic force prevails over the antagonistic force F1 of the return spring 82 of the armature.

After making the race S1, the armature comes into contact with the shutter, which is subjected to the biasing force of its return spring 91 and the hydraulic forces.

This results in a sharp increase in the mechanical counterforce, which can then exceed the magnetic force.

But with the increase of the pressure compensation effect under the surface of the shutter resting on its seat, the counterforce decreases again, and towards the end of the lifting movement there is again a margin excess of magnetic stress. As already indicated several times, we choose the value of the mass of the shutter, to ensure a rapid disappearance of the twists due to the shock of the armature and the shutter arriving against each other in opposite directions; ; It is necessary that the mechanical force at the end of the opening of the valve is greater than half the saturation magnetic force, to ensure a rapid return movement and of short duration, the shutter twists at the end of the movement of the valve. recall of it stop quickly, because the return spring acts on 11 shutter with a fairly high effort, while the closing speed is low.

 The application of the features according to the invention will now be explained in the case of injection valves already known as to their general configuration.

 FIG. 7 shows an electromagnetic injection valve whose armature is spherical and whose magnetic circuit is formed of laminated elements made of thin strip. The skirt 106 of the magnetic circuit is constituted by a series of fingers which separate openings of large area.

 The armature 113 is guided by the sheet of the skirt, with a slight radial clearance. Thanks to this magnetic circuit with a low loss field, it is possible to use a frame of reduced size and low mass, without the electromagnetic efficiency being much reduced. A thin washer 105 of non-magnetizable plastic material is mounted between the upper plate of the yoke and the skirt, to define the residual air gap. The upper plate 104 of the yoke is threaded onto the core 101. Inside the frame 108 of the winding, an elastic sheath 107 of plastic material is fixed by gluing or welding, so as to provide a cavity between the winding and this sheath, in order to damp the hydrodynamic oscillations.In the interior of the armature, the additional mass 110 is mounted. This additional mass 110, associated with the main return spring 103 and the return spring 111 which is weak, has the effect of producing a sudden variation of the opening effort. To reduce the reluctance, the pole of the core 101 is adjusted to fit the profile of the spherical frame, with a narrow flange to avoid the effect of hydraulic bonding. This collar is only a few hundredths of a millimeter high, to allow a quick balancing of the pressure under it. support face of the pole. A fuel circulation is maintained in the injection valve, to prevent the formation of bubbles.

 To reduce twists and mechanical fatigue of the seat of the valve, the body 114 of the injector is made with a thin wall. Here again, the natural frequency of the body of the injector is tuned to quickly stop the rebound of the armature 113 which is due to the impact between the moving parts in opposite directions. The plane of the assembly joint of the housing is arranged in the vicinity of the pole, in order to avoid centering problems. The travel of the armature can be adjusted by rotating the core by means of a screw system; the mechanical force is adjusted at the end of the stroke by turning the adjusting screw 100.

Figure 8 shows an electromagnetic injection valve having a spherical armature and an atomizer device. The magnetic circuit comprises a housing 120, a core 121 forced into the housing, a breech plate 127, and a spherical armature 126. The armature is guided in the breech plate, with a slight radial clearance, to ensure The breech plate 127 is securely attached to the body 128 of the injector, non-magnetizable material, for example by gluing, crimping or welding. At the same time, the reinforcing plate, integral with the body of the injector is centered by a flange, to ensure proper centering of the armature.For the damping of the hydrodynamic oscillations, the frame 123 of the winding comprises a internal cavity, closed at its upper part by an annular seal 122 of non-magnetizable material and electrically insulating. This annular seal is fixed by gluing or welding. This cavity can also be made directly during manufacture of the mounting frame, for example by blowing or by a similar method,
The pole's pole is spherical, with a radius much greater than that of the spherical armature, with a difference of a few hundredths of a millimeter. Thus, the cross section of the air gap tends to widen from the inside to the outside, which avoids a hydraulic bonding, and ensures effective damping of the armature at the end of its lifting stroke. The difference in radius also makes it possible to compensate for slight imperfections in the centering of the core. The fuel that reaches the shutter passes almost entirely through small holes in the cylinder plate 127. Depending on the flow velocity of the fuel, a significant throttling effect occurs in these holes, resulting in a significant depression at full opening of the valve. This depression produces in the direction of closure a return force which depends on the flow. Even at low opening of the valve, according to the ratio of the seat diameters of the valve and the armature, a considerable return force is obtained; and if the diameter of the ball is sufficiently strong relative to that of the seat of the valve, it results in a stress curve with an upward slope accentuated, for an increasing opening of the valve.This reaction effort is fairly accurate in a mass production, even with relatively inaccurate manufacture of the fuel inlets. In this way, it is possible to dispense with a particular adjustment of the return force. The steep upward slope of the curve makes it possible to obtain a large return force at the end of the opening stroke, in a manner that is favorable to the operation of the valve, by providing the latter with a return movement of short duration. A throttling effect can also be achieved with radial grooves in the cylinder head plate, giving these grooves an oblique orientation to impart angular movement to the fuel. Obviously, it is more expensive to achieve these grooves with the desired accuracy than to drill simple holes. In addition, these grooves reduce the strength of the bolt plate, and the guiding accuracy of the armature.

 The rebounds that occur at the end of the lifting movement are reduced to a large extent by the hydraulic damping obtained in the impact indentation. With respect to the mechanical force at the end of the stroke, the force of the return spring 125 is low and serves only to ensure that the shutter is properly sealed on its seat, in particular when the engine is stopped.

 The cross section of the pole piece of the core 121 is much stronger than that of the rest of the core. Thus, despite the small thickness of the wall of the core, a saturation speed is obtained a magnetic force barely exceeding the mechanical stress at the end of the race. Thanks to this arrangement, the inductance of the winding is increased for the same number of turns, and therefore the thermal fatigue is reduced. It is possible to use a very simple control circuit without reducing the holding current. The reduction of intensity which is then always necessary, is obtained by means of an external resistance connected in series.

To obtain a fairly short booster time, it is always necessary to provide a residual air gap-if uncomplicated control circuits are used. This residual air gap is located between the cylinder head plate 127 and the housing 120. At the same time, this residual air gap is used to let the atomizing air pass. The atomizing air stream comes from a separate air filter (not shown) which is attached directly to the valve housing. The atomizing air passes through large openings in the casing, and at the same time serves to cool the electromagnet's elongation. Then the air passes through radial holes, which may also have orientation component in the tangential direction, to introduce a rotation effect.air then reaches the mixing chamber, located in the lower body 128.de injector. The intimate mixing of the fuel and the atomizing air occurs in the mixing tube 129. This comprises an internal passage which tapers in the direction of flow, to improve the effect of atomization of fuel in the air that travels at a subsonic speed. The atomization of the fuel is furthermore favored by an acute profile outlet spout formed at the end of the mixing tube.
The useful stroke of the valve can be adjusted by rotating the nozzle body. The body of the nozzle is then blocked once this adjustment is completed, preferably with a needle, to immobilize the body of the nozzle relative to the housing.

 FIG. 9 shows an electromagnetic injection valve provided with a hot air atomizer system. The magnetic circuit includes a thin-walled core 142, forced into a housing 141 of non-magnetizable material. The skirt 144 of the magnetic circuit is perforated, with large openings, and nested on a peripheral edge of the lower base plate 148. The additional mass 146 rests on an annular in-plane projection of the tubular core 142, and urged in support by a spring 143, so as to cause a sudden change in the effective force at the opening of the valve, in combination with the return spring 150.The tubular armature 149 is made with an extremely thin wall and with a large internal diameter , in order to ensure a low throttling effect at the fuel passage, and a low rate of eddy current losses. The frame has a collar that significantly improves its strength. In addition, this flange is located between the bottom yoke plate 148 and the tubular core 142, to ensure a compact embodiment of the magnetic circuit, and a partial magnetic shielding of the working gap, to help reduce the field of view. losses.

The armature, the guide tube, and the valve obturator form a single piece, the tubular portion of which is associated with the magnetic circuit has a wall thickness of only about 0.5 mm, whereas the wall of the The guide tube is only about 0.2 mm. As a result, the weight of the armature is less than 1 g, and the electrodynamic losses are reduced to a minimum, the injection valve considered can operate at a very low speed. very high maneuverability, with very low power consumption. Preferably, the reinforcement has a width of 5 to 8 mm. This large diameter of the frame makes it possible to have a valve seat of large diameter, whereby a high value fuel flow is obtained for a small stroke. of the armat-ure. The polar surface of the armature-has radial grooves to ensure the balancing of pressure in the high position of the armature. The bearing surface of the reinforcement or that of the core is ground in the radial direction, to avoid hydraulic bonding. Large diameter holes are provided at the base of the frame and in the area of the moving equipment, to facilitate the flow of fuel with a small throttling effect and pressure variation.

 The armature 149 mounted in the lower portion 151 of the housing. has a lower part and an upper part. The short contact lengths of the mobile equipment reduce the effects of friction. The body 152 of the nozzle, in the form of a washer, is forced into the housing, and has a low natural frequency.

The nozzle body and the seat hole can be machined in a single operation without having to change the clamping.

 The race of the armature is adjusted by modifying the position of the core 142. Then, force is pressed the adjusting finger 140 in the housing 141, to adjust the mechanical stress end of stroke. As the core and the adjusting finger have the same diameter manufacturing is simplified.

 To evacuate the heat, fresh fuel is constantly circulated in the injection valve. The fuel flows through large holes, which may have a tangential orientation to cause vortex flow. Then, the fuel arrives opposite the seat of the valve, to flow after that into the housing, crossing the armature and out passing between the core and the adjusting finger. It is therefore unnecessary to provide radial holes in these parts.

 The atomizer device is forced into the housing base. An insulating skirt 153, in low thermal conductivity material, provides protection against heat. The atomizer device comprises a tubular mount 154 for the mixing tube 155. The latter comprises at the top a fastening collar forcibly engaged in the tubular mount. The hot air for atomization arrives via a connector 156 which opens into the tubular frame of the mixing tube. The flow of hot air thus circulates by mounting around the mixing tube, in the opposite direction of the atomized fuel jet, to an annular nozzle disposed externally at the top of the mixing tube. Thus the mixing tube undergoes intense heating, ensuring the evaporation of the fuel which could tend to condense inside the tube. Near its outlet end, the mixing tube carries oblique guiding projections, to ensure the centering the mixing tube, and impart a vortex effect to the atomizing air stream. The hot air stream exiting the annular nozzle thus forms a vortex around the jet of fuel which is projected to spray in the center of the hot air vortex where there is a favorable depression to ensure the acceleration of the droplets of fuel. This results in a supercritical pressure ratio between the atomizing air and the pressure prevailing in the suction pipe of the engine, and this causes compressive shocks which improve the effect of omission.

To conclude, we give below some indications on the embodiment of the fuel pump and on the electrical control system thereof,
To operate the injection valve according to the invention, by creating the necessary pressure for the fuel, a feed pump is required. For low value feed pressures, various kinds of pumps are known. The regulation of the fuel supply pressure can then be carried out, in a known manner, by the simple effect of discharging the excess fuel. However, particular problems are encountered in the case of pumps intended to supply the valve of injection in accordance with the invention, under a pressure of the order of 1000 bar.A Because of this pressure it is necessary to use only a piston pump, the power to provide coach this pump is very high, and to reduce the power consumption, it is therefore desirable that the volumetric flow of the pump is not higher than necessary at each operating speed of the engine concerned. For example, the pump piston can be controlled by means of an adjustable eccentric, but the power absorbed by such an eccentric pump has a high hysteresis, and a direct adjustment device comprising a jack for acting on a control lever of the eccentric produces unacceptable reactions on the pressure thus obtained. In addition, the levered transmissions present difficulties, because of the high value to be provided for the transmission ratio, and the extremely high forces which it is necessary to apply to the lever, which is why it is desirable to provide an indirect adjustment of the pump. Usually, it is sufficient to have a pump with a single piston and to dispense with accumulator, the accumulation effect being thus ensured thanks to the compressibility of the fuel and the elasticity of the pipes.

 FIG. 10 shows the diagram of a control circuit of an indirectly adjustable fuel pump. An auxiliary pump first sends the substantially constant pressure fuel into an accumulator, from where the fuel exits to pass through a regulating valve before reaching the high pressure pump. The pressure of the auxiliary supply pump can be adjusted simply by discharging the excess fuel. The volumetric flow rate of the high pressure pump is adjustable by means of a low pressure cylinder. The pressure of the high pressure pump acts on the pressure regulator; it comprises a piston subjected on one side to the high pressure and on the other side to the force of an adjustable counter-spring.

Thus, the value of the setpoint pressure corresponds to the piston stroke of the regulator. Preferably, an opposing spring constituted by groups of elastic cups is used, because a high ratio between the force involved and the displacement of the piston is thus obtained. Under the action of the regulator, the low pressure cylinder is sometimes connected to the discharge tank, and sometimes to the auxiliary supply pump. To create a hysteresis effect in order to avoid the difficulties due to oscillations, it is possible to provide an overlap between the positions of the regulator. It is convenient to have the output of the regulator in the vicinity of the high vacuum circuit, so that the regulator also serves as a safety valve in case of malfunction of the pump.

 In the injection valve according to the invention, the kinetic energy of the armature serves to cause the opening of the injection needle, and the time that elapses, between the moment when the The current of excitation, and the moment when the movement of the needle starts, depends to a large extent on the order of magnitude of the voltage of the excitation current. To avoid the additional expense of an electronic circuit responsive to the fluctuations of this voltage, it is preferable to stabilize this voltage by an electronic device. As we do not use the normal power of a switching transistor under the voltage of 12 V which is usual on a vehicle, and that it is necessary to further reduce the intensity, it is interesting to increase the control voltage beyond this usual value of 12 V, up to a value of preferably between 60 and 100 V. To increase the voltage, an electronic transformer is needed, usually with a transducer. In order to control electromagnets having a low rate of eddy current losses, the costs to be expected for the components can be substantially reduced by eliminating the transducer, the winding of the electromagnet of which can then be reduced. The accumulated energy for excitation can then be discharged between the excitation phases via one or more diodes to charge a reserve capacitor. We will expose the operating mode of such a circuit, with reference to Figure lla.

FIG. 11a shows a control circuit for actuating two electromagnetic injection valves
M1 and M2. However, this circuit is also suitable for controlling any number of injection valves, provided that the operating phases of these different valves do not overlap. The circuit has a high capacitance capacitor CL. When the excitation current is cut off on the electromagnet of a valve, the capacitor CL charges under the effect of the energy of the electromagnetic excitation field, at a voltage greater than that of the on-board circuit. vehicle in question A Zener ZD diode is provided to limit the voltage in case of malfunction of the circuit. The capacitor is connected in series with the supply voltage supplied by the on-board circuit, so that for the The excitation of the electromagnets, the edge voltage and that of the capacitor, are added.For the sake of clarity, the control circuit has not been represented.The mode of operation of the circuit under consideration is explained below. in the case where it is necessary to ensure the cyclic operation of the electromagnet M1, it is assumed that the charge capacitor has already been charged to the full value of its operating voltage.

 At the beginning of the excitation period of the electromagnet M1, the two transistors T1 and T2 are closed at the same time2 so that the electromagnet receives a voltage equal to the sum of the voltage of the on-board circuit and the charge voltage of the capacitor. The diode D1 prevents a short circuit of the capacitor. Thanks to the high value of the voltage that it receives, the electromagnet is excited quickly under a relatively low intensity. This phase is called the fast excitation phase. Towards the end of this fast excitation phase, the transistor T1 is turned off. The sustaining intensity which is then necessary is minced by modulating the supply intensity supplied by the edge circuit, through the diode D1. During the current breaking phases of the transistor T3, a slow or rapid drop in the excitation intensity can be obtained. A fast drop is obtained by cutting the transistor T2, which ensures a sending of energy to the charge capacitor through the diodes D1 and D2. When the transistor T2 is closed, the electromagnet is co-circled through the diode D3, which causes a very slow current drop, without sending energy to the charge capacitor. Thus, it is easy to adjust the voltage of the charge capacitor by acting on the transistor T2, to close or cut it, preferably during the sustain current phases. In addition, the proposed circuit provides a great deal of freedom as to a change in the pace of the excitation current during the lifting of the injection needle, and thereafter.

 When it comes to operating with short injection times and the circuit has not yet worked, it may happen that there is not enough energy to charge the capacitor. In such a case, the excitation of the electromagnet occurs before or after each working phase thereof, with a chopping of the excitation current, just sufficient for the magnetic force not to exceed the current. It is then possible to transmit sufficient energy, even if the antagonistic force is low, because the magnetic force increases according to a quadratic law, and the gap is important in the low position of the armature. It is also possible to provide an additional energy and the transmitted energy by prior excitation of the electromagnet.

 To study the shape of the current which is established to control the electromagnet, it is obviously necessary to mount resistance sensors in the circuit. They have not been represented for clarity. To vary. the injection process, it can provide a setting of the charging voltage. In particular for low pressure injection valves, this adjustment can be performed in an integrated circuit, associated with the logic trigger circuit, which makes it possible to dispense with external power transistors, thanks to a good use of the voltages of the transistors. the output stage of the proposed circuit. In addition, this circuit has a high security. even in case of malfunctions, because the intensity is always limited by the winding of 1 electromagnet.

The stability of the lifting process can be further improved by coupling the energy of the magnetic field obtained by discharging the capacitor. This discharge can be effected during a half-oscillation, but this requires control circuits. It is possible to have particularly simple circuits if this energy transmission is carried out in only one quarter oscillation, as in the case of the circuit of FIG. 11b. In this circuit, the trip logic is very inexpensive, and the entire circuit is particularly suitable for controlling high-pressure injection valves.
In the circuit of FIG. 11b, a capacitor CL with a relatively low capacitance is used. The energy accumulated in this capacitor varies in linear function of its capacitance, and as a function of the square of the charging voltage. The value of the charging voltage is chosen to ensure the accumulation of a sufficient amount of energy for a capacity preferably comprised between 2 and 10 microfarads. This requires a relatively high charging voltage of between about 100 and 300 V, depending on the size of the injection valve. For a given value of the duration of lifting and a given value for the inductance of the electromagnet, the value of capacitance of the capacitor is chosen to have a minimum energy consumption.

 From an external source of current, the capacitor is charged to the voltage UH. In principle, it is possible to use as a voltage source an oscillator of the so-called blocking type or of the non-blocking type. In a non-blocking oscillator, the energy transmission takes place during the flow phase of the transducer. Theoretical study shows that even when charging the capacitor at the ideal efficiency of the oscillator, it is impossible to to exceed a yield of 50% with respect to the energy thus sent to the capacitor, because of the considerable loss due to the internal resistance. On the contrary, a blocking oscillator makes it possible to charge the capacitor with a low rate of loss, because the energy from the magnetic field of the transducer passes during the blocking phase, and the electromagnets provide constant pulses of energy, regardless of the charging voltage. Thus, in the case considered, it is appropriate to use as current sources only voltage transformers operating according to the principle of a blocking oscillator. The maximum value of the charging voltage of the capacitor is limited by an electronic device by cutting off the power supply. It is interesting to be able to vary the charge voltage to modify the injection process.

 The circuit of FIG. 11b can be operated with any number of electromagnets, at will, provided that their operating phases do not overlap.

We will expose the operating mode of a system where it is a question of controlling the electromagnet M4. The discharge of the capacitor is triggered by the simultaneous closing of thyristor Th and transistor T3. The winding of the electromagnet and the capacitor then constitute a resonant circuit. The thyristor Th included in this circuit has the function of switching, after reaching the maximum intensity or when the voltage goes through zero, to prevent a current return in the winding and a negative charge of the capacitor.In addition, by isolating In this way, the charge capacitor can be used again even during the active phase of the electromagnet, and a large number of electromagnetic injection valves can thus be controlled in this manner by means of blocking oscillators. weak power.

 With small blocking oscillators, there is usually no need to turn off the power because the corresponding energy is too low to prevent switching. And it is therefore sufficient to have a single regulating device for the maximum load voltage, which operates independently of the injection phases of the valves thus served. But it is also possible to replace the thyristor with a diode, with a much lower useful life for the charge of the capacitor between the injection maneuvers, so that a maximum power blocking oscillator is needed. strong. However, this blocking oscillator can be used as needed to produce the holding current.

 Diode D5 prevents short circuits with the on-board power supply. After the blocking of the thyristor, the power supply continues with the edge current at voltage UB. In the circuit considered, the direct supply from the on-board circuit produces an exponential drop in intensity, at a slow rate, but with the injection valves in accordance with the invention, which have a low rate of current losses. Foucault and a high return force, this fall does not cause an unacceptable return delay when the duration of injection is brief. With windings whose resistance is low it is necessary to provide in series with the diode D5 a resistor which limits the intensity of the holding current. Better yet, it is possible to provide a circuit for regulating this intensity. On the other hand, in the presence of less stringent requirements with respect to dynamics and with relatively low return forces, the diode D5 can be directly connected to the Instead of connecting it to the on-board voltage, the holding current is then produced by the magnetic field of the solenoid. However, if the dynamic requirements are severe, it is always desirable to additional stabilizing device for the holding current, or for the supply voltage. To ensure a rapid reduction of the field after a fast excitation phase, the transistor is briefly cut at the end of the lifting process. To limit the extra-breaking current, it is necessary to provide additional protection, of course, although It can also limit the holding current by practicing modulated cuts. The known circuits of this kind are easy to combine with the circuit according to the invention, and it is therefore not necessary to give a detailed description. Obviously, when modulated cuts of the holding current are used, the fidelity of the injection process is slightly affected, since the electrodynamic conditions then vary a little in the return phase, depending on the direction of variation of the maintaining intensity, growing or decreasing, at the time of complete cut-off.

 In conclusion, it should be made clear that the arrangements provided for in the invention are not only applicable to electromagnetic injection valves of the kind under consideration. Indeed, the arrangements according to the invention can be used in all cases where it is to achieve very fast return movements, with great fidelity and very low energy consumption. In addition, the proposed injection valves can be used in a slightly modified form, to realize fast electro-valves for fluid circuits and in particular any hydraulic circuits. In particular one can also use the electromagnets already proposed by the Applicant in DE-A 3,314,900, by providing for the large surface openings proposed above, and by reinforcing the elements of the magnetic circuits by ribs or fallen edges. These magnetic circuits can be provided with enlarged polar faces and / or flanges.

 On the other hand the elementary parts of the electromagnetic injection valves can be made by manufacturing them in various ways; for example, the electromagnet parts can be made by frittarge, stamping, lap forming, or machining with chip removal.

 It is possible to use a hydraulic return effect in almost all known injection valves of the low pressure type having a radially guided armature. For this purpose, it is sufficient to ensure a throttling of the fuel flow inside the valve, between the top and bottom of the moving parts, in order to obtain a return force as a function of the flow rate.

Claims (1)

 CLAIM  Method for increasing the detonation limit in a spark-ignition engine, comprising a device for injecting water into the suction pipe, characterized in that the injection of water is carried out only at large regimes engine power, and especially according to the indications of a detonation detector.