g”
`
`-
`HPBooks 30;;"3‘3;
`'41."):12:43}:
`
`
`
`BOSCH ImmummWmmum:g
`'°
`
`
`
`
`FIIII. INJE“ION SYSTEMS
`
`
`HING‘-j
`
`
`
`
`
`
`
`VW EX1009
`
`US. Patent No. 6,557,540
`
`VW EX1009
`U.S. Patent No. 6,557,540
`
`

`

`BOSCH
`
`FUEL
`
`INJECTION
`
`SYSTEMS
`
`Forbes Aird
`
`HPBooks
`
`

`

`HPBooks
`
`are published by
`The Berkley Publishing Group
`A division of Penguin Putnam Inc.
`375 Hudson Street
`
`New York, New York 10014
`
`First edition: July 2001
`ISBN: 1—55788-365-3
`
`© 2001 Forbes Aird
`
`10 9 8 7 6 5 4 3 2 1
`
`This book has been catalogued with the Library of Congress
`
`Book design and production by Michael Lutfy.
`Cover design by Bird Studios
`Interior illustrations courtesy of Bosch, Inc. and the author as noted
`
`All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form,
`by any means electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the
`publisher.
`
`NOTICE: The information in this book is true and complete to the best of our knowledge. All recommendations on parts
`and procedures are made without any guarantees on the part of the author or the publisher. Author and publisher disclaim
`all liability incurred in connection with the use of this information. Although many of the illustrations in the pages that
`follow were supplied by Bosch and used with their permission, this publication is a wholly independent publication of
`HPBooks.
`
`

`

`CONTENTS
`
`CHAPTER 1
`Food For Engines, Food For Thought
`
`CHAPTER 2
`Fuel Injection: Then and Now
`
`CHAPTER 3
`
`Bosch Intermittent Electronic FI
`
`CHAPTER 4
`
`Motronic Engine Management
`
`CHAPTER 5
`
`Troubleshooting Bosch Intermittent Electronic FI
`
`CHAPTER 6
`
`Bosch Continuous Injection
`
`CHAPTER 7
`Troubleshooting Bosch Continous Injection
`
`CHAPTER 8
`
`Performance Modifications
`
`19
`
`37
`
`59
`
`75
`
`85
`
`109
`
`123
`
`

`

`This material may be protected by Copyright law (Title 17 U.S. Code)
`
`

`

`
`
`
`
`
`
`
`
`
`
`
`
`
`Bosch Fuel Injection Systems
`
`
`
`THE REAL IERK
`
`In a diesel, the cylinder at the end of the compression stroke is filled with nothing but air; combustion cannot
`
`begin until the fuel is injected. Thus, the moment of introduction of the fuel is equivalent to the ignition timing on
`
`a spark-ignition gasoline engine. For that reason, control is needed over not just the quantity of fuel injected, but
`also the timing of that injection.
`
`Because the fuel has to be kept separate from the air until the correct instant for ignition, the fuel is injected direct-
`
`ly into the combustion Chamberche injector nozzle lies inside the cylinder. As a result of that, the injector is
`
`exposed to full cylinder pressure during the power stroke. To prevent combustion pressure from blowing the fuel
`
`backward through the fuel lines, the injector is fitted with a very stiff spring-loaded check valve. This check valve
`
`also performs a number of other important functions. It prevents fuel from dribbling out of the injectors between
`
`timed squirts, and ensures that the start and end of each injection is clearly defined. Without this valve, the fuel
`
`Spray would taper off gradually toward the end of each injection. The check valve, in conjunction with similar
`
`check valves at the pump outlets, also ensures that the fuel lines leading from the injection pump remain filled with
`
`fuel at full pressure, to ensure that the pressure pulse from the pump is transmitted immediately to the injector noz-
`
`zle. To overcome the resistance of these check valves, the entire system has to operate at a very high pressure—as
`much as 3000psi!
`The design of the classic diesel injection pump, and the gasoline injection pumps directly derived from that
`design, follows from the three requirements: to control the quantity of the fuel delivery, and its timing, and to pro-
`
`
`vide a very high pressure output. In construction, it consists of an approximately rectangular body, bored with a
`
`
`number of small cylinders—one for each engine cylinder—each of which is lined with a cylindrical sleeve. The
`
`
`whole thing thus roughly resembles a miniature engine block, a similarity that extends to the "bottom end" of the
`
`
`pump, which has a shaft running lengthwise through it, analogous to an engine's crankshaft. Instead of crank
`
`
`throws, hOWever, this shaft has a number of individual cams, one for each cylinder. Each cylinder in the pump con-
`
`
`tains a piston—like plunger that is driven up its sleeve by the corresponding cam lobe as the pump shaft rotates, driven
`
`
`by the engine at half crank speed, like an ordinary ignition distributor.
`
`
`The mechanical drive thus provides the timing and the necessary pressure, leaving the problem of control of fuel
`
`
`quantity. To achieve this, each sleeve has a "spill" port drilled through its side, while the upper edge of each plunger
`
`
`has its side cut away to form a sort of Spiral ramp. The sleeve fits in its bore with a very slight clearance, and so is
`
`
`free to rotate. Rotation of the sleeve in its bore thus covers or exposes the spill port, according to the angular posi~
`
`
`tion of the sleeve relative to the spiral cutaway on the plunger. With the sleeve rotated so that the spill port is cov—
`
`
`ered even when the plunger is at the bottom of its stroke, the only route out of the cylinder is through a hole at the
`
`
`top, where the fuel lines to the injectors connect. A single stroke of the piston will thus expel the entire volume of
`
`
`fuel held in the cylinder; this represents the maximum capacity of the pump and thus corresponds to full power.
`
`
`At anything less than wide-open throttle, a lesser amount of fuel per revolution is obviously needed, and this
`
`
`reduction is achieved by rotating the sleeve around, relative to the plunger, so that the spill port lies some distance
`
`
`above the plunger's spirally formed top edge. Upward movement of the plunger thus initially causes fuel to be dis—
`
`
`charged from this port, from where it is fed back to the tank, until the plunger rises far enough to close off the port,
`
`
`and injection commences.
`
`The rotation of the sleeve(s) is accomplished by a toothed rack that meshes with gear teeth cut onto the outside
`
`
`of each sleeve. The rack is connected directly to the throttle linkage, so moving the pedal rotates the sleeves with-
`
`
`in the pump, thus controlling the quantity of fuel delivered with each stroke of the plunger.
`
`
`Although such a mechanical injection pump is, in fact, quite a simple device, and contains few parts ~just two
`
`
`per engine cylinder served, plus the camshaft and rack—the quantity of fuel delivered per stroke is very small, so
`
`
`all those parts have to be made with extreme precision, involving much hardening, grinding and precision gaug-
`
`
`
`
`
`
`20
`
`

`

`Fuel Injection: Then and Now
`
`ing. Accordingly, mechanical injection pumps—known widely and for fairly obvious reasons as "jerk” pumps—
`are mighty expensive. When adapted for use with gasoline, rather than the kerosene used in diesels, there is the
`additional problem that gasoline is a "dry" fuel, lacking the lubricating properties of diesel fuel, thus demanding
`the use of extremely hard, wear—resistant alloys, which further adds to the machining difficulty and expense.
`
`Mercedes-Benz SL 1954
`
`The famous Mercedes 3DOSL "gull-wing" coupe introduced fuel injection to the street. Like the W196, a system of direct injection was
`used. (Mercedes Benz).
`
`inverted flight, most of the advantages listed
`are also highly applicable to race cars.
`Accordingly, by 1937 Mercedes Benz had
`tested a single cylinder mockup for a racing
`engine, using high-pressure injection of fuel
`directly into the combustion chamber. In the
`meantime, a handful of European makers of
`small,
`two-stroke engined passenger cars
`had introduced this "direct" fuel injection, in
`an attempt to tame the notorious thirst of
`two—strokes that results from something like
`one quarter of the air/fuel mixture whistling
`straight out
`the exhaust ports during the
`intake/exhaust event, when both inlet
`
`("transfer") and exhaust ports are open
`
`together. Because these engines were two—
`strokes,
`the injection pump had to run at
`crank speed, rather than half-speed, as for a
`four—stroke. This provided Bosch with valu-
`able lessons in running their injection equip-
`ment at high pump rpm. Mercedes's first
`post—war Formula One race engine made
`direct use of the experience so gained, both
`their own and that of Bosch.
`
`In both the M196 Formula One engine
`that appeared in 1952 and in the engine for
`the famous 3DOSL "gull-wing” sports car,
`first exhibited in New York in 1954,
`
`Mercedes retained the direct (into the com—
`bustion chamber) injection scheme used in.
`
`21
`
`

`

`Bosch Fuel Injection Systems
`
`
`
`The tocation of the injector—screwed right into the cylinder head—45 apparent in this cross—section of the 3ODSL engine. (Daimler-
`Chrysler Archive)
`
`diesels and in the earlier experiments with
`gasoline fuel. Combined with suitable injec—
`tion timing, this allowed radical valve tim-
`ing and experiments with "tuned" intake
`pipes, without introducing the problem of
`poor fuel economy from portions of the
`intake charge being lost out
`the exhaust.
`Until injection occurred, after the exhaust
`valve had closed, the engine was inhaling
`only air.
`
`Diesel Jerk Pump
`Given the existence of an established tech—
`
`nology for injecting fuel, it is hardly surpris—
`
`ing that the diesel "jer " pump was adapted
`in this way for gasoline injection. The adap—
`tation, however, required hurdling a major
`difficulty-the problem of regulating the
`quantity of fuel delivered throughout the full
`range of engine operating speeds and loads.
`Unlike gasoline engines where power is
`controlled by closing off ("throttling") the
`air supply, yet where a very narrow range of
`air/fuel
`ratios must be maintained, on a
`
`diesel there is no "throttle" as such. At any
`given speed, the engine inhales the same full
`load of fresh air no matter what the position
`of the gas pedal; varying the power is sim-
`
`22
`
`

`

`Fuel Injection: Then and Now
`
`
`
`The resemblance to Diesel engine practice is clear in this exterior shot of the 3005L engine. Gasoline's tower lubricity, compared to diesel
`fuel, compelled the use of expensive, wear-resistant alloys in the injection pump. (Daimler-Chrysler Archive)
`
`ply a mater of injecting more or less fuel.
`Power is regulated,
`in other words, by
`adjusting the mixture strength. In the sim—
`pler diesel systems, at least, this is directly
`controlled by the driver's right foot—the
`engine runs extremely lean at light throttle,
`and rich to the point that it visibly smokes
`when maximum power is demanded. In fact,
`it is the smoky exhaust that sets the limit for
`the rated power of a diesel engine; more
`power would be available simply by inject-
`ing yet more fueliif we were prepared to
`put up with the smoke. (Even at full power,
`the amount of fuel injected falls short of 3
`
`stoichiometric mixture—diesels are always
`running "lean," which is one reason they can
`never produce as much power as a gasoline
`engine of the same displacement.)
`The inefficient breathing of any piston
`engine, gas or diesel, at speeds well away
`from the rpm at which torque peaks means
`that the mass of air inhaled per revolution, at
`any given throttle setting, will most definite-
`ly not be constant across the speed range.
`Because of the gasoline engine's finicky
`appetite, the amount of fuel mixed in with
`that air also has to be varied on the basis of
`the quantity of air inhaled, and not just
`
`

`

`Bosch Fuel Injection Systems
`
`3. Sensor an contoured tom
`
`2 2. [antral rack lreatl
`3. Enrichment solenoid
`4. Thermostat
`5. Barometric tell
`6. Check valve
`
`7. Plunger unit
`8. loathed segment
`9. (antral rack
`
`16. Shut-off solenoid
`
`10. Roller toppet
`ll. Camshaft
`12. Governor control lever
`13. Contoured tom
`
`14. Centrifugal governor
`15. Idle adiusting screw
`
`The complexity of the mechanical controls for the Bosch system used on the 3005L is mind-boggling. Fuel delivery of the pump (only the
`camshaft is shown, 11) is controlled by the rack (9). The position of the rack is governed by engine speed, via the centrifugal weights
`(14) acting on a contoured cam (13), and by throttle position. Other factors are accounted for by a barometric capsule (5) and a tem-
`perature sensitive device (4). (Robert Bosch Corporation)
`
`according to speed and throttle position.
`Adapting to Gas Engines—To adapt a
`diesel jerk pump to gasoline operation, then,
`means that the rack (see sidebar, page 20)
`cannot simply be directly hooked up to the
`loud pedal; the relationship between the two
`has to be modulated by some other con—
`trol(s). The usual way this is done is by
`adding a set of centrifugal weights, some-
`what like those in a traditional ignition dis-
`tributor, to provide a signal proportional to
`engine speed, plus a diaphragm that "reads”
`
`manifold vacuum, a fairly close approxima—
`tion of engine load. These additional control
`devices are connected to the linkage control—
`ling the rack via a system of cams and links,
`with the shape of the cams tailored accord—
`ing to the idiosyncrasies of the engine's
`breathing.
`The monkey motion of these additional
`levels of control added even more to the
`
`substantial cost of the pump itself, demand—
`ed meticulous adjustment, and its complexi~
`ty mocked the essential simplicity of the
`
`24
`
`

`

`Fuel Injection: Then and Now
`
`basic pump. Indeed, one observer at the time
`mused aloud that if all engines had worn
`fuel
`injection all along,
`then someone
`invented the carburetor, he would have been
`hailed as a genius.
`from the
`Injector Location—Apart
`mechanical complexity of the "add-on"
`speed— and load—sensing mechanisms at the
`pump, another problem confronted this
`adaptation from diesel to gasoline—that of
`shielding the injector nozzles from the heat
`and fury of combustion. Surprising as it may
`seem, gasoline combustion chambers see
`higher peak temperatures than diesels. On
`their M196 Formula One engine, Mercedes
`dealt with this by fitting the injector into the
`side of the cylinder, where it was masked by
`the piston at TDC and thus shielded from the
`worst of the inferno.
`
`Certainly, timed direct injection deals with
`the potential
`issue of fuel consumption,
`especially with radical valve timing, and the
`high pressure spray of these systems is
`favorable for atomization, but against these
`advantages were set
`the problems noted
`above, plus the issue of the power required
`to drive a pump that has enough grunt to
`blow open the stiff check valves in the injec—
`tors and at the pump outlets, used respec—
`tively to prevent combustion pressure from
`blowing back through the injection lines and
`to
`keep
`the
`lines
`fully
`charged.
`Nevertheless, the fact that engines operate
`quite happily on carburetors made it clear
`from the outset that there was no absolute
`
`need for precision timing of the fuel delivery
`in a spark—ignition engine.
`Granted, a carbureted engine would prob-
`ably make a little less power than a similar
`one with direct injection, but some part of
`this might be attributed to the fact that a sep»
`arate injector for each cylinder eliminates
`the problems of a central mixer, rather than
`to the timing. Of course, the same thing can
`be achieved with carburetors if a separate
`carb is provided for each cylinder. That,
`however, would be heavier, probably more
`
`expensive and arguably even more complex,
`depending on the number of cylinders. Thus,
`while a handful of other Formula One teams
`
`and upmarket European sports car manufac-
`turers also used Bosch gasoline injection, to
`the best of this writer's knowledge, all these
`others left the injector just outside the com-
`bustion chamber, aimed at, and spraying
`against, the head of the intake valve.
`But if you decide to locate the injectors
`outside the combustion chamber, then you
`
`don't need to fight combustion pressure, so
`you don‘t need a very high pressure pump.
`Also, if the injector lies outside the cylinder,
`then you don't need to time the injection so
`it occurs with both valves closed. In that
`
`case, why bother to time it at all? And if the
`tinting of discrete squirts is not crucial, then
`why does the fuel have to be delivered in
`individual pulses?
`
`Early Continuous
`Injection Systems
`These questions had occurred to a good
`many people, and an alternate form of fuel
`injection came into being. In these, fuel is
`injected continuously, rather than in pulses,
`and at very much lower pressure than a
`"jer " pump provides,
`though still much
`higher than the small pressure difference
`that exists between a carburetor venturi and
`
`the atmosphere.
`The Wright brothers used a primitive form
`of this continuous injection, as the arrange-
`ment is called, on the engine of their first
`successful aircraft. An engine—driven pump
`of the ”positive displacement" typeeone
`which supplies a fixed quantity of fuel at
`each revolution—was used, to continuously
`spray fuel directly into the engine air intake.
`It should be clear that while a satisfactory
`mixture strength might be arranged at full
`throttle—presumably after some tinkering
`with the size andfor drive ratio for the
`
`pump—with no means to reduce the pump
`output the mixture must have become hope—
`lessly rich at anything less than full throttle.
`
`25
`
`

`

`Bosch Fuel Injection Systems
`
`
`
`Although Stu Hilborn had proved, years earlier, that such a system could work on the track, Rochester were the first to demonstrate the
`practicality of the simpler, continuous flow fuel injection system, as on this early "fuelie" Corvette. (Dave Emanuel)
`
`then the Wrights' had no immediate
`But
`interest in operation at part throttle!
`Winfield’s FI System—About the same
`time that Bosch and Mercedes were begin-
`ning their experiments with timed, high-
`pressure, direct gasoline injection in the
`mid-1930s, Ed Winfield developed a low
`pressure, continuous
`injection system,
`intended for Indy cars. Winfield's design had
`provision for varying the fuel flow with
`throttle position, but a lack of interest from
`potential customers forced him to eventual-
`ly abandon the scheme, and he allowed his
`patents to lapse. It was not until fifty years
`after the Wrights that racers had access to a
`continuous injection system that dealt with
`the part—throttle problem.
`Hilbom Injectioantu Hilborn achieved
`this by incorporating a controllable spill
`valve, connected to the throttle linkage, that
`
`tapped off (and returned to the tank) some
`portion of the pumps output, according to
`how wide the throttle was opened.
`On the face of it, this might appear to fill
`the bill, but the assumption built-in here is
`that an engine, at any given throttle setting,
`needs a fixed quantity of fuel per crank rev—
`olution, and we have already seen that this is
`mistaken, because of the way an engine’s
`ability to take a full breath varies across its
`speed range. As a result, gasoline Fl systems
`that use just engine speed and throttle posi—
`tion to determine fuel quantity can be
`arranged to provide a suitable mixture
`strength only under very limited circum—
`stances, and must inevitably miss the mark
`by a large margin under others
`Still, schemes such as Hilborn's have a
`
`long and honorable history when used with
`methanol fuel. As discussed in more detail
`
`26
`
`
`
`

`

`Fuel Injection: Then and Now
`
`in Chapter 8, methanol will tolerate wide
`variations in mixture strength with little
`effect on power output, which helps mask
`the inherent shortcomings of such a simple
`system. In this case, rpm and throttle posi—
`tion alone provide a sufficiently accurate
`measure of fuel flow requirements, so it is
`not necessary to measure the rate of airflow
`at all. Since there is no need to measure it,
`there is no need for the restriction of a ven-
`
`turi or any other airflow sensor stuck in the
`airstream. This improves engine breathing
`and so promises more power just for that
`reason alone.
`
`Rochester FI System
`For a street engine running on gasoline,
`however, any purely mechanical system,
`whether injecting continuously or in a series
`of pulses, needs to somehow measure or
`compute the rate of airflow into the engine,
`and to use that information to modulate the
`
`delivery of fuel so as to keep the air/‘fuel
`ratio appropriate to the circumstances. We
`have discussed how this can be achieved
`
`with a jerk pump, but what about a continu-
`ous gasoline injection system?
`There have been many such systems
`throughout automotive history; one of the
`more recent and more widely used of these
`is the Bosch K~Jetronic system, described in
`Chapter 6. Nevertheless, the example most
`of us remember best is likely the Rochester
`system used 0n "fuelie" Corvettes from
`1957 to 1965. The heart of the Rochester
`
`system is a spill valve that serves as a regu-
`lator of fuel quantity—indeed, it is the only
`mixture control element in the whole sys-
`tem. Fuel is fed to the spill valve by an
`engine-driven gear-type pump that simply
`serves as a source of fuel at medium—high
`pressure (200psi at maximum), with the
`pressure tending to rise with increasing rpm.
`The spill valve consists of a sleeve with a
`number of ports arranged radially in its side,
`plus a plunger that covers or exposes these
`ports, according to its position. The position
`
`of the plunger, in turn, depends on a balance
`between the fuel pressure and a force
`applied to the top of the plunger by a mix-
`ture control mechanism.
`
`Increasing fuel pressure tends to raise the
`plunger,
`thus uncovering the sleeve ports
`and so bleeding off some of the pressure.
`The mixture control device, shown schemat—
`
`ically in the following illustration, amounts
`to a simple mechanical computer
`that
`applies a spring force to the top of the
`plunger, modified by the forces acting on
`two diaphragms—one connected to the
`intake manifold, and one to a venturi
`
`through which all the air entering the engine
`has to pass. The eight individual port injec-
`tors downstream of the spill valve thus see a
`fuel pressure that depends on a balance
`between the supply pressure and the forces
`acting on the diaphragms, and their output
`varies accordingly.
`The Rochester system is strikingly simple,
`especially when compared to a jerk pump
`having the controls needed for street use on
`gasoline, and none of its parts require excep-
`tional precision or expensive materials for
`their manufacture. Although the injectors
`themselves had to be individually calibrated
`and installed as a matched set, given suffi—
`ciently large production numbers it seems
`likely that the system overall would have
`been little (if any) more expensive to make
`than a pair of four barrel carburetors.
`Nevertheless, it remained a somewhat rare
`
`and costly performance option, and justified
`its price premium by offering more power.
`Partly this was because the near—elimination
`of cylinder—to—cylinder variations in mixture
`strength permitted a higher compression
`ratio (CR) than was safe with carburetors,
`but an advantage still showed even when the
`CR was unchanged. We might attribute this
`to three sources: nearly identical air/fuel
`ratios at each cylinder, elimination of mani-
`fold heating, and a reduction in the breathing
`restriction imposed by the venturi, com—
`pared with a carburetor. Although a venturi
`
`27
`
`

`

`Bosch Fuel Injection Systems
`
`MANIFOLD VACUUM
`
`
`
`'
`
`'tlllllllllO
`
`aa
`
`.
`
`OUTPUT TO
`
`INJECTORS
`
`The heart of the Rochester system's control unit was a spill valve that returned to the tank a certain fraction of the fuel delivered by the
`gear-type pump. The position of the piston in the spill valve was determined by two diaphragms—one hooked to a manifold vacuum, one
`to a single large metering venturi that fed the air box. A system of links and rollers linked movement of the diaphragms to the spill valve
`piston.
`
`the
`was needed to measure air quantity,
`amount of pressure drop needed to provide a
`useful signal to the mixture control unit was
`much less than that required to lift gasoline
`out of a carburetors discharge tube.
`In its day, the Rochester F1 was regarded
`by the auto industry and the public alike as
`very much a high-tech piece, and possibly
`the final word in fuel delivery. Ironically, at
`about the same time that GM was introduc—
`
`ing this system, another completely different
`form of fuel injection was being unveiled, to
`much less public acclaim, yet this one would
`
`eventually sweep away not only the carbu—
`retor, but most other forms of fuel injection
`as well.
`
`The First Use of Solenoid Valves
`
`The basis of this little-hailed breakthrough
`was the use of an electrically operated sole—
`noid valve as a fuel injector: Supplied with
`fuel at a modest constant pressure, it squirts
`at a constant rate as long as an electric cur-
`rent is applied to the solenoid windings, and
`stops when the current is turned off. Thus,
`the quantity of fuel injected depends simply
`
`28
`
`

`

`Fuel Injection: Then and Now
`
`on how long the juice is turned on. If one
`electrical pulse is delivered every second
`revolution of the crankshaft (for a four—
`stroke engine), then the air/fuel ratio of the
`mixture delivered to the engine can be con—
`trolled simply by varying the length of each
`
`pulse.
`The idea of using a solenoid valve as a fuel
`injector is a surprisingly old one. Such a
`scheme had been developed by an engineer
`named Kennedy at the Atlas Imperial diesel
`Engine Co,
`in 1932, and an installation
`appeared on a marine engine exhibited at the
`New York Motor Boat Show in 1933,
`around the same time that Bosch and
`
`Mercedes in Germany, and Winfield in the
`US. were beginning their experiments.
`Given the primitive state of the electrical
`sciences back then, it is mildly astonishing
`that the thing worked at all, yet in 1934 a
`similar but smaller engine was installed in a
`truck and driven successfully from Los
`Angeles to New York and back.
`The ”Electrojector"—A quarter century
`later, A.H. Winkler and R.W. Sutton of the
`
`Eclipse Machine Division, Bendix Aviation
`Corporation, announced the fruits of four
`years’ work at Bendix to develop amt Fl sys-
`tem for gasoline—fuelled automobiles using
`such a solenoid—controlled valve as an injec—
`tor, in conjunction with an electronic control
`unit, or (as they termed it) "brain box. " They
`described this combination of components
`as "electronic fuel
`injection"r~surely the
`first time this phrase was ever used. The
`resulting ”Bendix Electrojector" comprised
`a separate port injector for each cylinder,
`each supplied with fuel from a common rail
`at a regulated 20 psi. Sensors responsive to
`manifold pressure, engine speed, atmos-
`pheric pressure, and air and coolant temper—
`ature fed their various outputs to the "brain
`box"—the electronic control unit (ECU)-f
`which then calculated the instantaneous fuel
`
`requirements of the engine. On the basis of
`that calculation, the ECU delivered a pulse
`Of current to a rotary distributor which fed
`
`that current pulse to each solenoid vaive in
`turn, causing each successively to open and
`supply fuel to the cylinder it served. The
`quantity of fuel delivered by each injection
`was simply a matter of the duration of the
`pulse—as long as it was "on," the injector
`would deliver fuel at a constant rate. At idle,
`
`the pulses were of very short duration; at
`maximum load, the injectors were open for
`as much as 150 crankshaft degrees.
`At the beginning of the development of
`the Electrojector, about 1951 or 1952, the
`triggering signal for the ECU was provided
`by a mechanical commutator—a simple
`wiper making intermittent contact with a
`segmented brass ring, as the engine—driven
`ring swept past. That soon gave way to trig-
`gering by an extra set of contact points in a
`housing sandwiched under the ignition dis—
`tributor. In at least the three earliest versions,
`
`. vacu—
`.
`the working bits in the ECU were .
`um tubes! One such unit was installed on the
`
`V8 engine of a 1953 Buick, and used to
`demonstrate the system to the auto industry.
`By 1957, when it became available to the
`public in very limited numbers, the electron-
`ics were, thankfully, all transistorized. This
`eliminated the need to wait for the tubes to
`
`warm up before the system became func—
`tional, shrank the size of the ECU and sub—
`
`stantially reduced the electrical power
`required to run it, and doubtless greatly
`improved the system reliability—vacuum
`tubes do not take kindly to be rattled around!
`The Electrojector was listed by American
`Motors as an option for the 1957 Rambler
`Rebel, but very few cars so equipped were
`ever sold. A year later, Chrysler offered it on
`the 300D and DeSoto Adventurer, and sev—
`
`eral hundred with the option were manufac—
`tured in 1958 and 1959.
`
`Despite occasional problems from exter—
`nal electrical fields causing the brain box to
`go haywire, and a degree of unreliability of
`the injectors themselves, on the whole the
`system worked satisfactorily and matched
`Bendix's claims of modest
`(5 percent)
`
`29
`
`

`

`
`
`Bosch Fuel Injection Systems
`
`improvements in both fuel economy and
`power, compared to a similar engine fitted
`with a pair of four—barrel carburetors. The
`teething troubles were eventually worked
`out, and by 1965 Bendix had a fully devel-
`oped production version ready to go.
`At this point, rather than try to produce
`and market
`the Electrojector themselves,
`Bendix decided instead to enter into an
`
`agreement with the Robert Bosch Co.,
`which gave Bosch access to the Bendix
`patents. From a business point of view, this
`made a good deal of sense. In view of their
`reputation in the auto industry as an estab—
`lished OEM supplier—one with wide expe-
`rience both in fuel injection and in electrical
`and electronic equipment—Bosch was
`clearly in a better position to exploit the
`potential market. And fuel prices in Europe
`were many times what they were in the U.S.
`of the 1950s, justifying a much higher initial
`cost there, in exchange for the fuel economy
`benefits alone.
`
`Bosch Ietronic Fl
`In most respects, the Bosch "Jetronic" FI
`that first appeared on the 1967 Volkswagen
`1600, was functionally identical
`to the
`Electrojector system described by Winkler
`and Sutton ten years earlier. Apart from tak-
`ing advantage of advances in electronics to
`produce a more compact ECU,
`this sys-
`tem—since identified as D-Jetronic, to dis-
`
`tinguish it from later Jetronic versions — dif-
`fered from the Bendix prototype in one
`respect. In the Electrojector, the housing that
`was piggybacked onto the ignition distribu-
`tor and which contained the triggering con-
`tact points also contained a rotor, just like an
`ignition rotor;
`this distributed the timed
`pulse to each injector sequentially. In the D—
`Jetronic as applied to the four cylinder VW,
`there was no rotor, but there were two sets of
`triggering contacts, one for each pair of
`cylinders. Each set of contact points drove
`two injectors simultaneously, timed so that
`one cylinder received its injection while its
`
`intake valve was open; for the other cylin-
`der,
`the injection occurred with valves
`closed. Fuel thus accumulated in the port,
`awaiting that cylinder‘s next intake stroke.
`We have mentioned earlier that the mere
`
`fact that a carbureted engine will function in
`a satisfactory way should make it obvious
`that injection timing is, to say the least, not
`critical, and the power gains realized with a
`few continuous injection systems tended to
`confirm this view. The matter was pretty
`much clinched in 1959, when Mercedes suc-
`
`cessfully uSed a Bosch jerk pump with just
`two plungers, each feeding three engine
`cylinders, on its six cylinder 2208B. This
`must all have improved the comfort level of
`the Bosch and VW engineers who adopted
`the hop-skip timing used on the D—Jetronic,
`an arrangement that allowed eliminating the
`rotor used in the Electrojector.
`intermittent
`The Electrojector—based,
`injection ”Jetronic" systems from Bosch
`have undergone successive improvements
`and refinements over the years since the D—
`Jetronic's introduction. Apart from the elec—
`tronic innards of the control unit gaining
`computing power while shrinking ever fur—
`ther in size and weight, and the use of all—
`electronic triggering (rather than contact
`points), the major differences between one
`variant and another have been in the way the
`quantity of air supplied to the engine is mea—
`sured. These basic differences between sys-
`tems are reflected in the nomenclature used:
`
`L-Jetronic, LH-Jetronic, etc.
`In addition,
`most later systems accept inputs from addi—
`tional sensors, most notably an oxygen sen—
`sor, also called an 02 sensor or lambda sen—
`sor. As mentioned in the previous chapter,
`this is a device fitted to the exhaust system
`tha "sniffs" the exhaust gas to gauge its
`oxygen content, thereby providing a direct
`measure of what the air/fuel ratio actually is
`at any given moment, rather than obliging
`the system to depend totally on a set of pre-
`programmed 1‘maps" or "look-up tables"
`that pred