AIR-ASSISTED FUEL DIFFUSION

The subject of the invention concerns a device for the injection of fuel into th suction pipe of combustion motors. The device belongs to the category o injection systems with air-assisted fuel diffusion. The fuel is rπetered by mean of one or more electromagnetic valves and is then transported with the assistanc of air to the individual injection locations via small diameter conduits. I addition, a suitable air pump for the generation of pressurized air is described which functions in conjunction with the fuel injection device and is drive directly by the combustion motor itself.

Objective of the Invention and State of the Art

The basic design of a device for fuel injection with air-assisted fuel diffusio is familiar, for instance from DE-OS 2920636 (Em enthal et. al.). In th referenced application no detailed descriptions for constructioin of the injectio device are made.

A practically useful design example of such a device is described i US-Patent 4,708,117. The electromagnetic valve of this device features a fla valve seat with an annular channel. By means of the annular channel very goo precision of fuel distribution to the individual cylinders of the combustion moto is achieved. However, the electromagnetic metering valve must be built with high degree of precision, causing difficulties in manufacturing. Despite the hig precision requirements in manufacturing, the sealing capacity of the valve seat i often unsatisfactory. Furthermore, the speed of the opening and closi movements, and their reproducibility, is not always adequate for the desire

metering precision in the case of this valve. Alternatively, the use of a ball-type valve obturator has been proposed. With the use of this type of obturator it becomes necessary to arrange the individual metering nozzles in close proximity to each other. This in turn leads to mutual interaction between the flows to the individual nozzles. This type of nozzle arrangement below a ball-type valve obturator results usually in unsatisfactory precision of fuel distribution.

It is furthermore state of the art, to assign an individual injection valve to each cylinder of the combustion motor, and to equip these injectors with devices for the air—assisted fuel diffusion. The injection valves are then usually supplied by means of a fuel manifold which is located above the injection valves. This results in high costs and a rather undesirable additional equipment height above the individual valves. In addition, mounting such an injectio system is complicated due to the fact that a multiplicity of individual element and connections results. Because of the relatively large distance of th injection nozzles, strong pressure variations arise inside the individual valves, and, because of the impaired perfusion of the individual valves with fresh fuel hot-start problems occur frequently.

Air consumption of this type of injection device with air-assisted fue diffusion decreases with increasing amount of injected fuel. This behaviour i attributable to the fact that, as the amount of injected material increases, fue progressively blocks the supply lines, impeding the air flow. This featur considerably complicates adaptation of the rated performance of an air pump to th rated requirements of the injection device.

For state of the art systems, an electrically driven air pump is used t generate the compressed air. The electric drive for the air pump causes increas manufacturing costs. Therefore, it has been desirable to use an air pump which

directly mechanically driven by the combustion motor itself. The use of s pumps has already been proposed several times. The state of the art pum however, feature an extraordinarily unfavourable response curve for the amount air delivered. With directly driven mechanical pumps, the amount of air delive increases approximately linearly with the number of revolutions, on the ot hand, the amount of air required by the injection device decreases with increas number of revolutions. With state of the art pumps it is therefore necessary employ complicated and costly means to limit their output.

Applicant has proposed an injection valve with tilt-armature in the Ger Application P 3834445.9; this device allows for extraordinarily fast float movements and is easy to manufacture. This injector valve, however, is sadd with the disadvantage that the flat tilt-armature can only be adapted to the us requirements for rotationally symmetrical mounting conditions by more complica manufacturing measures. It is the objective of this invention to further deve this injection valve, featuring a tilt-armature, to a simple fuel injection dev which allows for an especially simple adaptation to the mounting requirements the enginge, despite the appearance of the initially unfavourable shape of t valve. The injection device shall make use of the known advantages air-assisted fuel diffusion, and shall simultaenously supply several cylinders a combustion engine with fuel. The fuel injection device shall allow especially cost efficient, uncomplicated manufacture, together with impro linearity and repeatability of fuel metering. The compressed air for f transport shall be produced by a simple, mechanically driven air pump.

Fuel Injection Device according to this Invention

The injection device according to the invention consists of a multifunctio

H composite in which an electromagnetic metering valve, a fuel pressure regulator, and several mixing chambers are integrated to one mounting part for the purpose of fuel delivery. The electromagnetic valve features a flat tilt-armature which can seal several valve seats. An alternative design can also be equipped with several individual valves. Inside the individual valve seats a separate single metering nozzle is provided, respectively coordinated to each individual cylinder of the combustion engine. The injection device comprises fuel conduits which serve for injection of the fuel into the suction tube of the combustion engine. A preferred design of the injection device is shown in cross-sectional view in Fig. la. The armature of the electromagnetic valve and the membrane of the pressure regulator are located in the same plane. Fuel supply, fuel recycle, and compressed air supply are defined by the directional arrows. Fig. lb provides a top view of valve carrier 102. The cross-sectional drawing shown in Fig. la corresponds to the axis A-A shown in Fig. lb. Identical parts are marked by the same reference numbers. The valve shown in Fig. 1 features four separate metering systems to supply a four-cylinder motor with fuel. The external dimensions of the injection device, in the design shown, are 70x40x40 mm. The injection device will be further detailed in the following, based on Figs, la and lb :

The housing of the injection device consists of valve carrier 102 and housi cover 101. Housing cover 101 is connected to valve carrier 102 by thread connections. The screws are located at the outer periphery of cover 101 and f into threaded holes 130. Membrane 119 is clamped between valve carrier 102 housing cover 101. The thickness of membrane 119 is preferably about 0.5 Membrane 119 consists of elastic material and serves to seal the inject device. In addition, membrane 119 provides the bearing for obturator 120 of fuel pressure regulator. Pressurized fuel flows via intake connector 103 to magnetic valve region. From there, the fuel proceeds via passage 107 to pressure regulator. Exiting from the pressure regulator, fuel flows via connec 104 to fuel recycle.

The fuel pressure regulator is in principle a state of the art device, and designed as a differential pressure controller. By means of such different pressure regulators, fuel pressure is maintained to a nearly constant amount ab the compressed air pressure, even as the compressed air pressure varies. Thus, metering nozzles 122 a virtually constant differential fuel pressure effective. The differential fuel pressure preferably is approximately 1 bar. compressed air pressure is widely variable, depending on the deliv characteristics of the air pump employed, the pressure in the induction tube the motor, and the amount of injected fuel. Compressed air pressure usua exceeds atmospheric pressure by several 0.1 bar in the case of an idling engi Under full load, compressed air pressure may exceed atmospheric pressure by much as 2 bar. Membrane 119, in the pressure regulation region, is impacted the fuel pressure on its upper side, and by the compressed air pressure on lower side. As differential fuel pressure increases, obturator 120 is lif against the forces exerted by governing spring 117 and the air pressure compressed air in . exit connector 104. This causes freeing of an annular s between obturator 120 and exit connector 104, allowing the pressurized fuel

I? pass to the virtually uπpressured fuel recycle section. Obturator 120 of th pressure regulator is supported by spring bearing 121. Spring 117 of the pressur regulator is borne by pin 116. Pin 116 is pressure fitted into valve carrier 102 The desired differential fuel pressure is set by the depth of insertion of pi 116. In place of the simple obturator described here, gimbal mounted obturators which are conventional to pressure control technology, may also be used.

Compressed air enters via connector 118 into the air space of the pressur regulator. From there it passes via passages 115 to mixing chambers 124. Th conduits for fuel mixture, 136, are inserted into mixing chambers 124. Transpor conduits 136 are provided with tapered entry sections 138 and are held in place b mounting clip 135. Gaskets 137 provide the necessary seal. The fuel-air mixtur is conducted via conduits 136 to the injection locations in the induction tube o the engine. The internal diameter of conduits 136 should be 1.5-2 times th diameter of the nozzles. Thus, the internal diameter of conduits 136 is abo 0.8-1.2 mm. Conduits 136 are preferably made of plastic materials. The diamet of mixing chambers 124 and passages 115 is preferably about 3 mm.

The magnetic circuit of the valve is defined by magnet core 113, fl armature 108, and valve carrier 102. All these components consist low-retentivity magnetic substances. Magnetic excitiation is by means of magnet coil 112, which is coiled onto coil body 111. Electrical connection is via conta pins 114. Armature 108 may be provided with a non-magnetic coating on i underside to establish a permanent air gap. By means of such permanent air ga the reset-time of the magnetic valve is shortened; a state of the art feature.

Flat armature 108 is in the shape of a tilt-lever with fulcrum on beari element 133. Bearing element 133 is arranged between fitting pins 109. Fitti pins 109 are pressure fitted into valve carrier 102 and provide lateral guidan to armature 108. The facing surface of magnetic core 113 is reset from the pla

of the armature bearing. Armature stroke 7corresponds to the measure of reset of the facing surface. Bearing element 133 should be located at such distance fro magnetic core 113 and valve seats 123, that a mechanical advantage with increase armature stroke results in the valve seat region. Armature stroke in the valv seat region preferably is about 0.1-0.15 mm, while the stroke at the magnet cor is preferably 0.05-0.1 mm. Armature thickness is preferably 0.8-1 mm, armatur mass is about 2 g.

Armature reset is by means of reset-spring 110. Reset-spring 116 is borne b calibration pin 106. Calibration pin 106 is provided with an extension 145 whic prevents spring 110 from dropping off during mounting of the injection device During mounting, armature 108 is fixed by electrically exciting magnetic coil 112 so that the armature does not drop out. Assembly then proceeds by placing valv carrier 102 onto cover 101, followed by threading both together. To provide fo easier assembly, cover 101 may alternatively be extended downwards close t armature 108, with the reset-spring then located in a drilled hole in cover 101 Under these conditions falling out of armature 108 and reset-spring 110 i prevented by the extended cover 101. Dynamic calibration of the valve is done b state of the art procedures involving the depth of insertion of the calibratio pin.

Armature 108 is additionally supported by limit stop 141 at the bearin location 140 of reset-spring 141. Limit stop 141 prevents unacceptable bending o armature 108 under the influence of the force of the reset-spring. Valve seat 123 are directly machined into valve carrier 102. Valve seat diameter i preferably about 2 mm. Bearing element 133, limit stop 141 and valve seats 12 are all located in the same plane. Between these parts, damping pockets 146 an 147 are arranged, which are preferably fashioned together by stamping of the valv carrier. Alternatively, such damping pockets may also be provided in armatur

108. For the given armature design, the depth of the damping pockets should be about 0.2 mm. This relatively large depth of the damping pockets is necessary because of the large fuel displacing surface of armature 108. The function and manufacture of the damping pockets is described in detail in German Patent Application P 3834447 (Electromagnetic Fuel Injector and Methods for its Manufacture).

Armature 108 is preferably manufactured by stamping. Armature 108 should be provided with thin lamellar extensions 132 in the region of valve seats 123. These thin lamellae should have a thickness of about 01.-02 mm. This provides for a certain elasticity of the armature in the valve seat region, and improves the sealing capacity of the valve seats. Alternatively, armature 108 may also be equipped with thin plastic obturators to improve sealing capacity. A further alternative would provide for arranging the individual metering nozzles within one single extended valve seat. With such an arrangement of several metering nozzles inside a single valve seat, the danger of mutual interference of the flow to the individual nozzles arises. Therefore, the arrangement of several separate valve seats, as shown, in general is favoured with respect to hydraulic considerations. To reduce weight and hydraulic forces during armature movement, armature 108 is provided with perforation 131. Armature 108 can additionally be provided with stamped ribs, or may be flanged at the outside or at the perforation, in all cases measures to improve its flexural strength.

The injection device according to the invention offers a multiplicity of essenti advantages in comparison with state of the- ar- .4ev.ices: To start, there is t advantage of simplicity of construction, where tolerances essential to- t functioning of the device can be maintained by simple production methods wi extreme precision. Finishing of the bearing plane of valve carrier 102 preferably achieved by a simple surface grinding process. This is followed stamping the damping pockets and valve seats into the flat surface of val carrier 102. Valve carrier 102 may also be executed as a sintered part, allowi for omission of the grinding and stamping steps, as long as the production of t sintered piece is done with enough precision. Furthermore, it is possible fashion the individual seating regions by stamping them, providing for t possibility of centering the stamping tool by means of a pin directly in t nozzle holes. This results in an exactly centered location of the valve seat wi respect to the nozzle opening. Such a method of production guarantees unifo in-flow conditions to the nozzle region. Final finishing of the bearing surfa of valve carrier 102 should always be done by flat-lapping. The final finishi step for the armature also is by means of a flat-lapping step.

By arranging membrane 119 in the dividing plane of the housing, it is n necessary to employ extensive sealing measures. The supply connectors can arranged freely in any location, providing for ready adaptability to changi mounting requirements. For example, air supply to the pressure regulator can arranged to be directly via bearing pin 116, which is then fashioned as connection piece. Design of the injection device in line with the insta invention provides for large and short cross-sectional diameters, with t pressure regulator being close to valve seats 123. This results in considerab reduction of hydraulic and pneumatic pressure variations, and decreases the dang of icing. All of the feed channels 115 to mixing chambers 124 can be execut with the same length. This guarantees identical oscillation and flow conditio

with respect to the individual metering nozzles. The pressure controller ca readily follow the pressure pulsations of the air pump, without this leading t unacceptable phase dislocations of the differential fuel pressure above th metering nozzles. Stable hydraulic conditions result, combined with very goo linearity of fuel metering.

In order to achieve the lowest possible amplitude of hydraulic pressur variations, the cross-section of connecting passage 107, between valve an controller, should be made as large as possible. Connection passage 107 should b executed to this effect as a short, broad channel, covering as much of the tota width of the controller as possible. The design of the injection device i keeping with the present invention allows for an extraordinarily larg cross-section of connection channel 107. In addition, the cross-section of fee connector 103 should be kept as small as possible in relation to the cross-sectio of 107. Favourable conditions are achieved for a feed connector cross-section

2 2 about 5 mm for 103, and a cross-section of about 100 mm for 107. By the measures, the formation of damaging pressure waves in the valve is considerab reduced, and their propagation into the feed conduits is largely prevented.

The injection device exhibits excellent characteristics with respect hot-starting. The complete fuel stream is supplied via the magnetic valve. T valve seats are located directly in the main fuel stream. This guarantees fre fuel being supplied to the valve seat region even under possible vapour-lo conditions.

With state of the art fuel injection devices, full fuel pressure maintained for several hours even after engine shutoff. This is necessary prevent evaporation of the fuel from the injection valve and thus to guarant fuel supply to the injector during a hot-start of the engine. To maintain f pressure a checkvalve is installed in the feed line to the injection valve; t

/ prevents fuel flow-back through the feed line. This involves tough requirement for the tightness of the valve seats, and the seat of the pressure regulator, t prevent untimely dropping off of the fuel pressure. In contrast, the excellen hot-starting ability of the injection device according to the instant inventio makes it possible to lower fuel pressure to atmospheric conditions immediatel after shutting of the engine. To this effect, a side stream line for fuel i connected parallel to the pressure regulator. To establish this by-pass stream return flow connector 104 is provided with a small drilled passage 105, it diameter should be of the order of 0.2 mm. The advantage of this lowering of th pressure lies in reduced tightness requirements for the valve seats and i improved working conditions for the fuel pump. Thus, during startup the fuel pum only works against an initially low pressure, guaranteeing the quick flushing ou of vapour bubbles. Even under unfavourable starting conditions, fresh fuel wil usually be supplied to the injection system within less than a second. Th desired by-pass can also be achieved by a rough grinding of the valve seat of th pressure regulator. This measure allows for a further reduction in costs.

Calibration of the injection device is done jointly for both the pressur regulator and the electromagnetic valve. To this effect, valve stroke should b chosen to be as high as possible so that at the individual valve seats fuel flo is choked only to the slightest degree. Static flow is set by setting th differential fuel pressure. Special calibration of armature stroke is no necessary. Dynamic flow characteristics are set by conventional means, by varyin the reset-spring force.

For the injection device according to the instant invention a canting of th valve obturators with respect to the valve seats is impossible, given the specifi design used. This makes for stable and uniform flow conditions in the individua seat regions, even so the armature stroke is very small.

Uniform fuel distribution is thus guaranteed, even for small armature strokes this in contrast to the state of the art injection devices of this type Therefore, in spite of the individual valve seats, the alternative exists t calibrate static flow via changing armature stroke height. Setting of th armature stroke is possible by respective changes in the depth of insertion o magnet core 113, either by pressure fitting or threading. Separate calibration o the controller is then not necessary.

The fuel injection device according to Fig.l is equipped with a simpl electromagnet where magnetic return flow occurs directly by means of th magnetizable valve housing 102. In the following two additional suitable magneti circuits are proposed for activating the valve; these can be installed alternatives in the injection device according to Fig.l, with location indicated by section line B-B.

Fig.2 describes an electromagnet with return flow via a yoke 204. T electromagnet is shown in cross-sectional view in Fig. 2a, in top view in Fig. 2 Identical parts are listed with the same reference numbers.

In fig. 2, the magnetic circuit consists of flat armature 202, magnet co 203, and the return flow part 204. These components are made of low retentivi material. The return flow part 204 takes the shape of a yoke and can manufactured as a stamped or sintered item. The yoke shape reduces the magnet resistance between the outer surface of magnet core 203 and the inner surface part 204 and thus reduces the stray field of the magnetic circuit. Magnet co 203 is pressure fitted into valve carrier 201. Valve carrier 201 should be made non-magnetizable materiel, since otherwise the magnetic circuit would be partial shorted out. Magnet core 203 can be provided with a longitudinal slot 207. T slot serves to prevent eddy currents and facilitates pressure fitting of mag core 203. Yoke 204 is clamped by collar 210 of magnet core 203. Magnet core

contains coil body 206, onto which coil 205 is wound. The center pole surface 20 should be of about the same magnitude as the total surface of the two side pole 209 togehter. As the armature is energized, it closes against the two side pol 209, while a permanent air gap of preferably about 0.05 mm remains between th armature and center pole surface 208. Alternatively, it is possible to arrange f two permanent air gaps in the region of the two side poles 209, a measure which less advantageous from magnetic considerations. On the other hand, yoke 204 largely relieved of mechanical stresses if the permanent air gaps are arranged the side poles. This has advantages from a vibration technical point of vie Final finishing of the pole surfaces should be done together with the val carrier after mounting.

Another suitable design for the magnetic circuit is shown in Fig. 3. In th case the circuit features two magnet cores 307 and 308, which are pressure fitt into valve carrier 301. The cores carry coil bodies 303 and 305, onto which coi 304 and 306 are wound. The always necessary permanent air gap is formed non-magnetizable valve carrier 301. Armature 302 closes directly against t magnet cores in the energized state. Coils 304 and 306 are excited together a are electrically in parallel or series connection. The coils are magnetized such a way that current flow in the two coils is in opposite directions. Th magnetic circuit design offers the advantage of simple and stable mechanic construction, combined with especially low stray field generation.

Fig. 4 shows an alternative mixing device which offers advantages in fuel-m generation, especially for very low pressure transport air. Conduit 408 in th case is provided with additional side-passages 406 and 407, these allow f additional air access from the side. This causes the fuel to foam up in t conduits at low air pressure. Such low pressures can arise during motor start when the injection device is driven by a mechanical pump, or when the injecti

device is provided with air at only atmospheric pressure.

The mixing device according to Fig. 4 is of similar design as that shown in Fig. 1. What is shown is a segment of valve carrier 102. Identical parts in both Figs. 1 and 4 are given the same reference numbers. Compressed air from the pressure controller enters mixing chamber 124 via passage 404. Passage 404 point toward injection nozzle 122, so that injection occurs about into the floodin point of the admitted air. This measure reduces the impact of pneumati pulsations on the linearity of etered fuel supply. Fuel-mix conductor 408 i inserted into mixing chamber 124. The upper segment of conduit 408 contains th tapered feed adaptor 411. Side stream air passes along the outside of 411 to sid passages 406 and 407. As fuel is injected, a low pressure zone develops insid 408, causing additional air to be sucked in through side passages 406 and 407 Conduit 408 is sealed by means of gasket 137 and is held in place by fastenin clip 135. Fuel-air mixture is carried by conduit 408 to the injection locations i the induction tube of the motor.

Supply of the injection device with pressurized air is preferably by means o a pump which is directly driven by the combustion engine. In general, the wel known positive displacement pumps are suitable. Such pumps are of the type o single or double—action membrane or piston pumps. By double-action pumps understand those pumps where two work spaces are arranged on either side of t displacement piston. For these state of the art positive displacement pumps, t suction side volume increases about linearly with the number of revolutions.

Fig. 5 shows a single-action air pump of the positive displacement typ Intake air enters via feed pipe 506, passes intake valve 503 and reaches cylind 501. Compressed air exits via exit valve 504 into outlet pipe 505. Outlet pipe 5 is connected to the air intake connector of the injection device. Cylinder 5 contains piston- 502 which goes through stroke s. The piston may be driven by

crankshaft or a cam. The upper dead center position of the piston stroke is sh by a broken line. At the upper dead center position of piston 502 rema distance s to the cylinder cover. The remaining cylinder volume is referred

as dead-space volume.

As the revolutions per minute of the combustion motor increase, the prob arises that the volume of compressed air produced grows considerably faster t the absorptive capacity of the injection device, if no restrictive measures taken. Because of the inadequate absorptive capacity, the compressed pressure, without restrictive measures, far exceeds the permissible pressure. rated volume efficiency of the usual pumps is totally unsuitable for use with injection device. In principle, there exists the possibility to limit compressed air pressure by means of a pressure regulator which is instal downstream from the pump. Excess air would then be bled off by the press controller. Such measures would, however, be burdened by a large drop in efficiency factor of the pump and would imply additional costs.

Applicants investigations have resulted in the knowledge that the famil positive displacement pumps can be adapted to the type of injection dev described in the invention; adaptation is possible by means of special variati in the dimensions of these pumps. Adaptation of the rated pump performance to adsorptive capacity of the injection device is by means of a significan enlarged dead space volume for the pump. The efficiency factor of such adaptation is considerably more favourable than that of controlled bleeding-off.

In addition, the larger dead space volume allows for an especially simple and cost efficient construction of the pump. The maximum air pressure produced decreases with increasing dead space volume. Dead space volume is increased to such an extent that the maximum pressure, under full load by the combustion engine, does not exceed about 2 bar. The dead space volume of a respective pump cylinder in general will be about 50% of total piston displacement. For conventional air pumps, the usual objective is to keep dead space to a minimum in order to obtain a pump with a high efficiency rating. Normally, dead space is reckoned to be 5-10% of pump displacement. By means of this considerable enlargement of dead space, in .contravention of the familiar basic rules of pump design, reliable pressure limitation is achieved.

In addition, feed connector 506 can be provided with a choke section 507; its free cross-section under stationary flow conditions should be 2-4 times the tota free cross-section of the fuel-mixture carrying conduits of the injection device. Exit pipe 505 can be connected with outside air via an additional intake pipe 50 and by-pass valve 508. This makes it possible to suck in additional air under th conditions where the pump, at lowest RPM's at lowest load of the engine, onl delivers low air volume; under these conditions a subpressure exists at the engin induction tube which activates by-pass inlet 508.

Furthermore, the pressure pulses generated by the pump should occu synchronous with the individual injection events. By synchronizing pressur pulses with individual injection events, the precision of fuel metering can b considerably improved. For trouble free performance of the injection device wit a four-cycle engine, at least one pressure pulse must occur for every two moto revolutions. Otherwise a variable, periodically rising and falling average rat of flow would result in the fuel-mix bearing conduits between sequential injectio events, this in turn would have the consequence of undesirable, periodical

rising and falling amounts of fuel injected. T Ahe pump is suitable driven by mea of a cam attached to the shaft of the combustion motor. This also synchroniz the pressure pulses of the pump with the individual injection events in t simplest possible way. Alternatively, a suitable pump can also be driven from t crankshaft of the combustion motor by gears or belts, either at the sa revolutions, or at half the revolutions of the combustion engine.

The amount of air aspirated by the pump per crankshaft revolution of t combustion engine, under idling conditions, should be about 1% of pist displacement of the combustion engine. For low RPM's, the air volume aspira per revolution of the pump corresponds approximately to the piston displacement the pump. By means of a correspondingly shaped double-cam, it is possible achieve one working cycle of the pump per revolution of the crankshaft. For pump working-cycle per revolution of the crankshaft, the pump displacement sho

? be about 20 cm ^" for a combustion engine of 2 Liter displacement.

As the combustion engine starts up, the problem exists that due to the v low starting RPM's, even with relatively large pumps, no significant excess sta pressure can be built up. The partial vacuum in the engine induction tube is a very low during startup and is not able to assist in a significant way generating an air stream in the fuel conduits. Despite these conditions, wit synchronized pump drive, starting of the motor is readily possible even given very small displacement volume of the pump. Initially, fuel piles up in the f conduits, more or less blocking them. Because of this blocking, pressure buil is then possible and the fuel is blown uout. This type of flow is referred to plug-flow. Fuel preparation can be improved in this case through the mix device described in Fig.4. In order to bring about the fastest possible press buildup, no pressure accumulator should be installed between pump and inject device. Such pressure accumulators are otherwise routine in pump construction

order to dampen pressure pulsations and ac Ihtieve uniform air flow. With the injection device according to the instant invention, omission of such a pressure accumulator does not cause any significant impairment of the precision of fuel metering.

Adaptation of the rated pump performance to the absorptive capacity of th injection device can be further improved by the solid state choke 507 in intak pipe 506. This results in additional limitation of the maximum flow volume. Ai flow at the choke location is strongly affected by pulsations. These pulsation become stronger, the closer the choke position is to the injection device Choke-pulsations can be decreased by interposing an accumulator between chok location and inlet valve. For an accumulator installed in this manner, resultin in uniform flow-through of the choke location, the free cross-section of th latter should be 2-4 times the total free cross-sections of all fuel-mix conduit of the injection device. If the choke location is very close to the intake valve its necessary free cross-section must be approximately doubled because of stron pulsations.

The injection device according to the instant invention can be equipped wit several electromagnetic valves so as to avoid an unfavourable excessive width an mass of the armatures. For instance, it is suitable for ό-cylinder motors, provide two separate electromagnetic valves, arranged in the same plane, whi respectively act on 3 valve seats. The valves are then electrically in a parall circuit. The electromagnetic valves can be arranged on both sides of the pressu regulator, allowing for the most compact design of the injection device. Wi several electromagnetic valves installed, it is possible to supply motors with to twelve cylinders from a single injection device.

It is also possible to arrange for several, separately trigger electromagnetic ^' valves in a common injection device, allowing for synchronizati

of the individual injections with the working c 7ycles of the respectively assign cylinders of the combustion engine. This type of operation is called sequenti injection. With this approach the formation of pollutants in the engine reduced, allowing for engine operation with a leaner mix. With several separa electromagnetic valves a highly cost-efficient injection device results which c satisfy high demands for engine control. Arrangement of the individual injecti devices directly inside a fuel manifold is epecially advantageous. Such a devi is shown in Fig. 6.

Fig. 6 shows a cross-section through a fuel injection device with separa valves, a cross-section through a single valve 604 is represented. The injecti valves are arranged next to each other directly in fuel manifold 601. Differi from the usual design, the injection valves are mounted without housing. T principal design of the individual injection valves has already been described German Application P 38 34 445.9. Injection valve 604 features a fl tilt-armature and is -fastened to the fuel manifold by two screws 609. In t example shown, the valve carrier of the injection valve is made as a sinter piece. At the clamping location the valve carrier features a neck, which serv to center valve 604 with repsect to fuel-mix conduit 603. The neck of valve 604 inserted into the common channel 611 of the fuel manifold 601. To seal 604, gasket 006 is provided. The magnetic circuit of valve 604 is located directly fuel channel 612. The magnetic coil of 604 is equipped with contact pins 61 Cabling of the individual valves is directly inside the fuel manifold 601. Th makes it possible to provide electrical contact for the valves by means of single central plug. In addition, the terminal electronic devices for triggeri the injection valves can also be arranged to be inside manifold 601. This resul in very good heat dissipation from the electronic parts, since they are in t fuel flow. No special cooling elements are necessary for the electron components. Triggering of the terminal electronic devices can be done with ve

0 small cross-section lines, further reducing the cabling requirements Air-manifold 602 is located below the injection valves. Air-manifold 60 intersects the individual mixing chambers, allowing for air supply to th individual conduits 603. The cross-section of air supply channel 602 should be a large as possible in order to reduce the amplitudes of pneumatic pressur variations. Manifold 601 preferably is made of light metal alloy or plastic Manifold 601 is closed with cover 607. Cover 607 is sealed by gasket 608 and i flanged onto manifold 601. Alternatively, cover 607 can also be made of plastic and can be joined to manifold 601 by means of ultrasonic welding. the alway necessary pressure controller can be built into manifold 601, or, as a separat commercially avilable item, can be joined directly to the manifold.

The injection device according to Fig. 6 posesses a multiplicity advantages in comparison with state of the art multi-point injection devices. start, there is the advantage of the especially simple and cost effecti manufacture. Due to the compound construction, a large number of individu connections are not necessary. This significantly simplifies final mounting the motor. The injection device can be supplied to the motor manufacturer as complete tested unit. There is little danger of improper handling a contamination by foreign materials. Compared to state of the art individu valves, considerably less danger of improper mounting exists during fin assembly. The plastic fuel-mix conduits 603 can neutralize even larger toleran deviations. In case of possible leakage at the joints of fuel-mix conduits 6 with manifold 601, only almost pure air will leak out, reducing the danger of fi hazard considerably in comparison with the conventional multi-point injectors. case of possible accidents, fire hazard is also considerably reduced.

In addition, the injection device features significant functional advantages, wh compared to the conventional multi-point injection devices with separate valve Only very minor hydraulic pressure pulsations occur, which significantly enhanc the precision of the metered amount of fuel, compared to the case for t conventional individual valves. Such pressure pulses form with the individu valves mainly in the relatively narrow connecting channels between fuel manifo and valve seats, and also inside the valve housing. Furthermore, inside t conventional individual valves very steep pressure peaks are generated by fu being displaced by the armature movement and also from housing vibrations. addition, in conventional individual valves the blowing out of vapour bubbl often is not readily accomplished because of the narrow connecting channels. the injection device according to Fig. 6, pressure pulsations are large prevented a priori by the total immersion of the valves in the fuel itself a because of the absence of a valve housing. The minor remaining pressure puls can dissipate very quickly because of the elasticity of the fuel conduits. For design where manifold 601 consists of plastic material, pressure pulses a virtually eliminated due to the elasticity of the material.

Due to the fact that the valves are completely surrounded by fuel, excelle hot-starting results , even for low fuel pressure. Therefore, it is qui acceptable to lower fuel pressure to atmospheric pressure directly after engin shutoff. Lowering fuel pressure after engine shutoff in the system according Fig. 1 is done by means of a by-pass which is parallel to the pressure regulator.

The conventional individual valves tend to the formation of deposits insi the injection nozzles. This tendency increases, with increasing thermal loadi of the injection nozzles. Thermal stress on the injection nozzles is exceeding low in the injection device according to Fig. 6, again because of the fact th the flowing fuel surrounds the valves, and because of the large distance of t

metering nozzles from the injection location. Thus, there is no danger here of deposit formation inside the injection nozzles.

The fuel injection device is mounted directly above the induction tube of the combustion motor. This results in only a short length of the individual fuel-mi conduits 603, usually in the order of 70 mm. Because of the short fuel-mi conduits, only a brief delay in fuel transport results. This results in very goo control of time sequential fuel delivery to the individual cylinders as th individual valves are triggered. Because of the compressed air diffusion of th fuel, excellent fuel presentation is provided. The injection device is wel suited to lean running of the motor.

With the low pressure difference of the transport air, the fuel can pas through the short, strait conduits (603) in a very short time, based on th kinetic energy of fuel stream alone. Therefore, even for very low differentia pressure of the transport air, a fast fuel supply to the motor is guaranteed Because of the short length of the transport conduits (603), the air supply lin (602) can be designed with a large cross-section in order to reduce pneumati pulsations. Air supply line (602) in this case acts as a pressure accumulator Without this feature, unacceptable amplitude phase dislocations of the pressur waves between the individual valves could occur. In contrast, a pressur accumulator of this type would neither be desirable nor suitable in a devic according to Fig. 1. In a device according to Fig. 1, where the metering nozzl and the pressure regulator are in close proximity, and given the symmetric design of the unit, there is no reason to fear any significant phase dislocatio Conduits (136) in the device according to Fig. 1 must be of significant lengt 300-400 mm, to establish connection of the device with the motor. Therefore, in device according to Fig. 1, omission of a pressure accumulator results in short fuel transport times under unfavourable running conditions, as has been describ

previously. ^3

Significant improvements in diffusion can be obtained by adding a speci spray-diffuser at the end of each transport conduit. To this effect, conduit 6 can be arranged to be inside an outer jacketing tube which is directly joined mixing chamber 611. This outer jacket tube takes a side stream of air from mixi chamber 611 and conducts it in the outer passage to the diffuser device which installed further downstream. The fuel-mix which was prepared in the mixi chamber passes then through the downstream diffuser device into the induction tu of the motor. The advantage of adding a downstream diffuser is to be found in t fact the the internal cross-section of the transport conduit can be ma relatively large in order to avoid wetting the walls of the conduit tubes wi fuel. Undesired wetting of the walls reduces fuel transport speed and impairs t quality of fuel diffusion. Wetting of the walls is caused by off-centered canted installation of the conduits with respect to the fuel stream exiting fr the metering nozzle. Under practical conditions, due to tolerance limits, fu contact with the wall can never be completely avoided, The limits which need to imposed on the amount of diffuser-air are set by means of a relatively sma cross-sectional flow diameter of the downstream diffuser device. T

2 cross-sectional flow diameter should be in the order of 1-2 mm . This side-str air is conducted through the jacket and fed concentrically past the outlet conduit 603 into the diffuser device. Any number of conventional diffuser desi can be used. The diffuser may for instance consist of a Laval nozzle or can be sharp-edged cone. In addition, conventional diffusers may also be directly joi to the transport conduits.

As is well known, fuel prepartion can be considerably improved by heating the mix component. With the device according to the instant invention, due to good heat dissipation through the fuel, and the total immersion of the inject

nozzles, no performance difficulties n Aeed be expected from vapour bubbl formation. Therefore, the use of familiar heating measures is especiall appropriate for the device according to the instant invention. For example, th injection device can be supplied with heated air, or even with off-gas from th motor. Heating the fuel-mix conduits is especially advantageous. Heating ma also be by electrical means, so as to add heat even during motor startup. To th effect electrical resistance heating wires can, for instance, be considered insi the conduits. Furthermore, by heating the transport conduits, the formation film-flow along the walls of the conduits can be prevented, even for very l differential pressure of the transport air. Heating allows for a considerab improved fuel atomization when the system is provided with atmospheric pressu air.

An especially simple design of an injection valve suitable for use with t device is shown in Fig. 7. Fig. 7a shows a cross-sectional view along line B- Fig 7b gives a top view of the valve as defined by line A-A. Identical parts a marked with the same reference numbers.

Valve carrier 705 of the injection valve consists of a single stamp sheeting part and is of low retentivity material. Valve carrier 705 features t side-brackets 708 which contain fastening holes 709. The valve is thread-connect at brackets 708 inside fuel manifold 712. To seal the valve at the cont location of valve carrier 705, a gasket ring 711 is provided. On valve rarr 705 is an upwards directed flange 704, which carries calibration screw 7 Calibration screw 701 serves to set the force of reset spring 702. The fo exerted by reset spring 702 should preferably be about 1-2 N. Reset spring forces armature 703 onto valve carrier 705. Armature 703 contains a non-magne pin 710 which is pressure fitted. Pin 710 is additionally fixed to armature by laser welding and serves as armature stop and as guidance for reset spring 7

Sidewise guidance for the armature is provided by the two pins 707. Armature 70 is also stamped with the impression 706 for improved flexural resistance. Magne coil 715 is on coil body 714, which is slipped onto the magnet pole, and als carries contact pins 713. The valve seat is machined into valve carrier 705. Valv carrier 705 is flat-finished. Armature 703, together with pressure fitted pi 710, is finished by grinding the bottom side, and the back edge of armature 703 i undercut by the height of the armature stroke. The permanent air gap is formed b a thin non-magnetic coating of armature 703 or valve carrier 705.

The special advantage of the valve according to Fig. 7 over conventiona valve designs is in the extraordinarily low manufacturing costs but, nevertheless improved functional performance. Armature 703 is very low in mass, amounting onl about to 0.3-0.4 g. Armature length is about 15 mm, width about 4 mm. Due to th rotational movement of the armature, only a small part of armature mass i involved with the complete stroke. Thus, the equivalent mass to an armature wit linear movement is only about 0.1 g. Conventional single valves, with a needl valve, in contrast have armature masses of from 2-4 g. The low armature mas allows for extraordinarily fast floating movements. Using conventional triggerin circuitry, on over-excitation of the valve during energizing, closing times o less than 0.3 ms can be achieved.

Therefore, due to the very high working speed of the valve, calibration of th reset spring is generally not necessary. Alternatively, the valve can also b equipped with a magnetic coil with very high electrical resistance, thus reducin the absorption of electrical energy. For example, the magnetic coil can have resistance of up to 60 Ohms, for more than 1000 windings, still allowing fo closing times of less than 1 ras. For such high coil resistance values, calibration of the reset spring should, however, be considered, as is proposed i Fig. 7.

Alternatively, the coil carrier may also be a composite of several parts The magnetic circuit can consist of a U-shaped segment, which is welded to t valve carrier along dotted line 720. The permanent air gap is ground into t magnet pole which is located at the junction shown by the dotted line. Thi first of all, results in the advantage that the valve carrier can be made non-magnetizable material. With a non-magnetizable valve carrier high attractive forces are generated, since in this case no magnetic forces acting the opposite direction to armature attraction are generated in the valve carri region. A further advantage results from the ability to oraitt the otherwi required non-magnetizable coatings. Also, bracket 704 can be executed as separate part. Bracket 704 can then be provided with extensions which are joint threaded with the the valve carrier. The advantage of such a design is to found in mechanical relief for the valve carrier. Calibration of the spring for of reset spring 702 can then be simply done by bending the bracket.

In conclusion, we wish to restate that, in place of compressed air, the fu injection device according to the instant invention can also be driven with air atmospheric pressure, or with off-gases from the combustion motor. The advanta of this resides in the lowered cost from the absence of the air pump. However, considerable impairment of fuel preparation must then be allowed for.

addition, with slight modifications, the device can also be used to advantage wi other injection procedures for fuel metering. For instance, the device can supp several injection valves in series arrangement which serve for the dire injection of fuel into the combustion area of the motor. The listed dimension as well as the methods for joining individual parts, are to be considered especially suitable, but only by way of examples. Pressure fitting may f instance be replaced by screw connections. Such simple modifications are readi apparent to those skilled in the art.

Additional suitable designs and variants of the fuel injection device can deduced from the claims.