Title:
ELECTRONIC COMPUTER FOR A SYSTEM OF FUEL INJECTION FOR COMBUSTION ENGINES
United States Patent 3863054
Abstract:
An electronic computer for a fuel injection system in which signals representative of the quantity of fuel to be injected are produced which are repectively related to a full load operating curve of the engine and to curve computed as a function of the real and theoretical engine speeds. The computer selects the signal representative of the smallest quantity and stores this to be used to produce the fuel injection signals.
US Patent References:
ELECTRONIC CONTROL SYSTEM FOR INTERNAL COMBUSTION ENGINES
Schlimme - January 1973 - 3707950
ELECTRONIC CONTROL SYSTEM FOR INTERNAL COMBUSTION ENGINES
Schlimme - January 1973 - 3707950
Application Number:
05/349718
Publication Date:
01/28/1975
Filing Date:
04/10/1973
View Patent Images:
Export Citation:
Assignee:
Societe des Procedes Modernes D'Injection Sopromi (Clichy, FR)
Primary Class:
Other Classes:
123/483, 702/47
International Classes:
F02D41/36; G06G7/64; F02D41/32; G06G7/00; F02M39/00; F02B3/00
Field of Search:
123/32EA,139E 235/150.21,150.2
Primary Examiner:
Botz, Eugene G.
Attorney, Agent or Firm:
Darby & Darby
Claims:
1. An electronic computer for computing the quantity of fuel to be injected for a fuel injection system for combustion motors, comprising:
2. A computer as in claim 1, wherein said signals produced by said first means are pulses and said first circuit means comprises a monostable circuit means delivering rectangular pulse signals of substantially constant duration having the same period as the pulses of the first means, and a non-linear filtering circuit responsive to said rectangular pulse signals for producing a voltage proportional to the speed of the motor.
3. A computer as in claim 1 wherein said second means for producing a signal corresponding to the posted speed of the motor comprises a potentiometer having its slider connected to a common potential point to divide the potentiometer into two branches, the electrical resistance of
4. A computer as in claim 1 wherein said signals produced by said first means are substantially rectangular pulses and are at a rate proportional to the speed of the motor and of a duration which is inversely
5. A computer as in claim 4 wherein said second circuit means includes a first monostable circuit means for producing a signal having a duration which is a function of the posted speeed and at the same rate as the signals produced by said first means, storage means, and fifth circuit means for producing a signal whose duration represents the difference between the signals produced by the first means and the first monostable circuit means, said fifth circuit means connected to said storage means for defining the charging time of said storage means, current generator means connected to said storage means for selectively producing a current which is a function of the posted speed to charge said storage means, the voltage stored by said storage means at the end of the duration of the charging time corresponding to the quantity of fuel to be injected
6. A computer as in claim 5 further comprising a second monostable circuit means connected to said first means for delivering a storage signal of fixed, short duration a storage circuit, transfer circuit means having non-linear filter characteristics connected between said storage means and said storage circuit to transfer the voltage from said storage means to
7. A computer as in claim 2 wherein said transfer circuit means having non-linear filter means comprises first and second transistors of opposite conductivity types, diode means for connecting the emitters of said transistors and diode means for connecting the bases of said transistors.
8. A computer as in claim 1 further comprising sixth circuit means for storing the voltage representative of the quantity of fuel to be injected, said sixth circuit means comprising:
9. A computer as in claim 8 wherein said storage circuit means comprises a capacitor and a transistor connected to said capacitor which receive said authorization signal, said capacitor being charged through said transistor in the absence of the said authorization signal to said voltage
10. A computer as in claim 8 wherein said discharge circuit means comprises a discharge resistance, a first transistor means connected to said storage circuit means and remaining blocked as long as the voltage level of the storage circuit means is higher than the said fixed threshold, second and third transistor means connected to said storage means made conductive by the appearance of the authorization signal and blocked by the change to the conductive state of said first transistor means, and a fourth transistor means connected to said second and third transistor means whose state is opposed to that of the two preceding, and which delivers the injection signal of duration proportional to the quantity to be injected.
11. A computer as in claim 1 further comprising means for generating an authorization signal to control the production of an injection signal comprising:
12. A computer as in claim 11 wherein the delay circuit means produces its signal after the appearance of the signal corresponding to the most remote possible moment of injection when the speed of the motor is below a
13. A computer as in claim 12 wherein said delay circuit comprises a capacitor, means for producing a current for a time equal to the delay imposed by said delay circuit means which is a function of the speed of the motor or the quantity of fuel to be injected to charge said capacitor from an initial charge voltage which is a function of the speed of the motor or of the quantity to be injected, up to a fixed charge voltage.
14. A computer as in claim 13 further comprising means responsive to the speed of the motor being below a certain threshold speed for at least
15. A computer as in claim 13 wherein the means for producing said charge current comprises transistor means whose control electrode receives the voltage proportional to the speed of the motor or to the quantity of fuel
16. A computer as in claim 13 further comprising discharge means, and means for diverting the charge current from said capacitor through said discharge means until the appearance of the signal corresponding to the most advanced possible time of injection, to insure the initial charge
17. A computer as in claim 11, wherein said means for producing a signal
18. An electornic computer and control device for a fuel injection system for combustion motors comprising:
19. A control device for transforming a voltage representative of the quantity of fuel to be injected into an injection signal of duration proportional to the quantity to be injected, comprising:
20. A control device for generating a signal for authorizing injections for fuel injection systems for a combustion motor comprising:
21. A computer as in claim 1 wherein said second circuit means operates to produce the voltage representative of the fuel to be injected according to the curves of statism in response to the voltages produced by said first
22. A computer as in claim 6 wherein said non-linear filter means comprises first and second transistors of opposite conductivity types, diode means for connecting the emitters of said transistors and diode means for connecting the bases of said transistors.
2. A computer as in claim 1, wherein said signals produced by said first means are pulses and said first circuit means comprises a monostable circuit means delivering rectangular pulse signals of substantially constant duration having the same period as the pulses of the first means, and a non-linear filtering circuit responsive to said rectangular pulse signals for producing a voltage proportional to the speed of the motor.
3. A computer as in claim 1 wherein said second means for producing a signal corresponding to the posted speed of the motor comprises a potentiometer having its slider connected to a common potential point to divide the potentiometer into two branches, the electrical resistance of
4. A computer as in claim 1 wherein said signals produced by said first means are substantially rectangular pulses and are at a rate proportional to the speed of the motor and of a duration which is inversely
5. A computer as in claim 4 wherein said second circuit means includes a first monostable circuit means for producing a signal having a duration which is a function of the posted speeed and at the same rate as the signals produced by said first means, storage means, and fifth circuit means for producing a signal whose duration represents the difference between the signals produced by the first means and the first monostable circuit means, said fifth circuit means connected to said storage means for defining the charging time of said storage means, current generator means connected to said storage means for selectively producing a current which is a function of the posted speed to charge said storage means, the voltage stored by said storage means at the end of the duration of the charging time corresponding to the quantity of fuel to be injected
6. A computer as in claim 5 further comprising a second monostable circuit means connected to said first means for delivering a storage signal of fixed, short duration a storage circuit, transfer circuit means having non-linear filter characteristics connected between said storage means and said storage circuit to transfer the voltage from said storage means to
7. A computer as in claim 2 wherein said transfer circuit means having non-linear filter means comprises first and second transistors of opposite conductivity types, diode means for connecting the emitters of said transistors and diode means for connecting the bases of said transistors.
8. A computer as in claim 1 further comprising sixth circuit means for storing the voltage representative of the quantity of fuel to be injected, said sixth circuit means comprising:
9. A computer as in claim 8 wherein said storage circuit means comprises a capacitor and a transistor connected to said capacitor which receive said authorization signal, said capacitor being charged through said transistor in the absence of the said authorization signal to said voltage
10. A computer as in claim 8 wherein said discharge circuit means comprises a discharge resistance, a first transistor means connected to said storage circuit means and remaining blocked as long as the voltage level of the storage circuit means is higher than the said fixed threshold, second and third transistor means connected to said storage means made conductive by the appearance of the authorization signal and blocked by the change to the conductive state of said first transistor means, and a fourth transistor means connected to said second and third transistor means whose state is opposed to that of the two preceding, and which delivers the injection signal of duration proportional to the quantity to be injected.
11. A computer as in claim 1 further comprising means for generating an authorization signal to control the production of an injection signal comprising:
12. A computer as in claim 11 wherein the delay circuit means produces its signal after the appearance of the signal corresponding to the most remote possible moment of injection when the speed of the motor is below a
13. A computer as in claim 12 wherein said delay circuit comprises a capacitor, means for producing a current for a time equal to the delay imposed by said delay circuit means which is a function of the speed of the motor or the quantity of fuel to be injected to charge said capacitor from an initial charge voltage which is a function of the speed of the motor or of the quantity to be injected, up to a fixed charge voltage.
14. A computer as in claim 13 further comprising means responsive to the speed of the motor being below a certain threshold speed for at least
15. A computer as in claim 13 wherein the means for producing said charge current comprises transistor means whose control electrode receives the voltage proportional to the speed of the motor or to the quantity of fuel
16. A computer as in claim 13 further comprising discharge means, and means for diverting the charge current from said capacitor through said discharge means until the appearance of the signal corresponding to the most advanced possible time of injection, to insure the initial charge
17. A computer as in claim 11, wherein said means for producing a signal
18. An electornic computer and control device for a fuel injection system for combustion motors comprising:
19. A control device for transforming a voltage representative of the quantity of fuel to be injected into an injection signal of duration proportional to the quantity to be injected, comprising:
20. A control device for generating a signal for authorizing injections for fuel injection systems for a combustion motor comprising:
21. A computer as in claim 1 wherein said second circuit means operates to produce the voltage representative of the fuel to be injected according to the curves of statism in response to the voltages produced by said first
22. A computer as in claim 6 wherein said non-linear filter means comprises first and second transistors of opposite conductivity types, diode means for connecting the emitters of said transistors and diode means for connecting the bases of said transistors.
Description:
The invention relates to an electronic computer for a fuel injection system for combustion engines.
The problem to be solved by a computer for a fuel injection system is essentially that of electronic determination of the quantity of fuel to be injected as a function of the speed of the motor, the posted speed and, optionally, other parameters. The term posted speed is used to designate the ordered speed, imposed for example by the use of the motor with the aid of a control organ, such as the accelerator. That is, for every position of the accelerator or other control organ there is a theoretical speed for the engine. It is known to electrically compute the quantity of fuel to be injected as a function of the speed of the motor and the posted speed. The signal produced by he computation is generally a voltage or a current, which is continually available. This type of computation is used with conventional fuel injection pumps to position the pump regulating mechanism.
It is also known, in the case of electronic fuel injection systems, about computing, whenever an injection is to take place, the duration of this injection. In other words, each time an injection is started or triggered, the computation of its duration is started by a measurement of all variables. The injection is stopped at the end of this duration.
It is also known to elaborate or compute and produce, before the injection, a signal representative of the duration of the injection, storing it in a suitable circuit, and "reading" it in the course of the injection so that the actual duration of the injection will correspond to that produced and stored.
The object of the invention is to provide a fuel injection duration computer which computes the injection duration information before the injection, and stores this information in a circuit for reading during the injection. The computer operates so that the computation is not integrally effected during the interval separating the injections, and the measurement of certain variables and certain preliminary calculations are effected on a continuous basis so that the intermediate results are available at all times.
The advantage of such a design is that it is thereby possible to make the measurements with greater precision and to process the data measured in such a way as to insure both precision and a sufficiently short response time. This is done while retaining the advantages of repeated computation, which are, primarily, lower sensitivity to electrical parasites, or interference signal, better stability, and the presence of periods when injection is authorized, the latter insuring security against eroneous injections.
The object of the invention is an electronic computer for a system of fuel injection for combustion engines of the type in which the electrical data corresponding to the quantity of fuel to be injected is determined before the injection and stored in a circuit for reading during the injection. The computer includes a first means for delivering pulses at a frequency proportional to the speed of the motor and a first circuit for transforming these pulses into an electrical voltage proportional to the speed of the motor. A second means is also provided to introduce a signal corresponding to the posted speed of the motor. A second circuit delivers after each pulse from the first means, a voltage representative of the quantity of fuel to be injected according to the "statism" curves. The second circuit produces its voltage by transforming the signals from the first and second means and, optionally, the voltage which is proportional to the speed of the motor. A third circuit also is provided which transforms the voltage proportional to the speed of the motor into a voltage representing the quantity of fuel to be injected according to a curve corresponding to the full load operation of the motor. Also provided is a fourth circuit which transmits the smallest of the two voltages representative of the quantities of fuel to be injected according to the curves of the statism (produced by the second circuit) and according to the curve of full load (produced by the third circuit).
Other characteristics and objects of the invention will appear in the description which follows, made in reference to the attached drawings wherein:
FIG. 1 is a group of curves representative of the quantity of fuel to be injected as a function of the speed for a compression-ignition motor;
FIG. 2 is a polygonal contour approaching the curve of full load C pc of FIG. 1, and capable of being computed and produced by a function generator;
FIG. 3 is a schematic block diagram of the system according to the invention, in which the continuous data is represented by dotted lines, the pulsed data by broken lines, and triggering signals by solid lines;
FIG. 4 is a diagram representing angular offsets of signals D and δ (delta), used for the production of certain electrical signals;
FIG. 5 is a timing diagram of the signals T o , T A , T S computed and produced by elements 1, 5 and 11 of the circuit of FIG. 3;
FIG. 6 is a timing diagram of distribution of signals D to T auto , computed and produced by elements 15 to 21 in FIG. 3;
FIG. 7 is a detailed diagram of elements 5 to 13 of FIG. 3, used for the computation and production of the voltage corresponding to the curves of statism;
FIG. 8 is a detailed diagram of elements 2 to 4 of FIG. 3, used for the computation and production of the voltage corresponding to the curve of full load; and
FIG. 9 is a detailed diagram of elements 17, 14 and 22 of FIG. 3, used for the computation and production of the injection duration.
In combustion motors, for example with fuel injection, the quantity q of fuel to be injected can be represented as a function of the speed of rotation N of the motor, by a group of curves such as that in FIG 1. In this figure, curve C pc represents the curve of full load for the motor, which can be exceeded only rarely. This complex curve can generally be approximated by a polygonal contour such as shown in FIG. 2.
The polygon of FIG. 2, corresponding to the full load curve, is varied or deformed as a function of the parameters of the motor (for example, supercharging with air at variable pressure), or for particular operating conditions (for example, starting). The elaboration, or generation, of the full load polygon of FIG. 2 is accomplished by an electronic function generator of any suitable known type.
In FIG. 1, the curves C s represent the curves of statism defined on the one hand by the posted speed of the motor N Ai , which is dictated by the user, and on the other hand by the function (dg/dN R) N Ai . The latter function depends, for a given curve, on the posted speed N Ai and on the real speed N R of the motor.
For a posted value N Ai of the speed, shown moving along one of the curves C s , and a real value of speed N R of the motor, shown by the dotten line, two values of the quantity of fuel to be injected can be determined. One is where the dotted line of N R intersects the curve of full load, at q pc and the other at point q s where it intersects the C s curve of statism N Ai .
The quantity of fuel to be actually injected is the smaller of two quantities q s and q pc . Naturally, if the real speed N R of the motor is higher than the posted speed N Ai , the quantity of fuel to be injected is zero.
In order to define the quantity of fuel to be injected, the real speed of the motor N R is measured as is the spread between the real and posted speeds (N R - N Ai ).
Depending upon the place at which the real speed is measured and the manner in which it is measured, the signal obtained contains an alternative or unwanted component due to an irregular movement of the motor component whose movement is monitored. An attempt can be made to eliminate this alternative component. However, this affects the response time of the measurement of the spread of the speeds.
Referring to FIG. 3, there is shown at 1, a fixed or static, means for delivering rectangular waveform signals of duration T o . The intervals T o of time corresponds to the time of passage before means 1, of a certain sector, of known dimensions, of a piece rotating synchronously with the motor, for example a sector on the fly wheel. This can be accomplished by a suitable electrical circuit or electromechanical switching means. It can also correspond to the interval of time separating the successive passages between two means, spaced by a known angular sector, of a reference mark rotating synchronously with the motor. This also can be done by suitable switching means as a photo-electric circuit. The interval of time T o is linked to the speed of rotation N R of the motor and to the angle A of opening of the sector considered. We have a relation of the type
T o = A/N R
the width (duration) of the rectangular signal T o being inversely proportional, and its frequency directly proportional to the speed of the motor. The T o signal is shown on the first line of FIG. 5.
The leading edge corresponding to the start of T o triggers a first monostable circuit 2 which delivers a rectangular waveform signal T v of constant duration. Since the speed of rotation of the motor is not always regular, but is subject to change, the period of signals T o , and the rate of production of signals T v is not exactly constant. The output signals T v of monostable circuit 2 is applied to a filter circuit 3 which transforms these signals into a voltage proportional to the speed N R . This voltage is also affected by a substantial spurious, or unwanted component. Filter 3 is therefore a non-linear filter which delivers a voltage V = λN, proportional to the mean speed of the motor. The voltage V is applied to a function generator 4 which delivers a voltage corresponding to the full load reading, and more exactly to the polygonal contour of FIG. 2.
The start of a T o signal also triggers a second monostable circuit 5 which produces a signal to determine a duration T A . The signal T A is a function of the value of the electrical signal representing the posted speed N A , defined below by α. The T A signal is shown on the second line of FIG. 5. In the interval of time (T o - T A ), as controlled by a circuit 6, and following signal T A , a capacitor 7 is charged linearly by a current generator 8 with a current i (α, N). The posted speed α is produced by circuits 9-10 furnishing two pieces of supplementary data: one, α to current generator 8, the other, (1- α 0), to monostable circuit 5. The current generator also receives the data of the duration (T o - T A ), or ΔT, and of the real speed of the motor V = λ N. The time ΔT is shown on the third line of FIG. 5. The level of voltage reached by the capacitor 7 represents the quantity of fuel q s to be injected according to the curve of statism C s (FIG. 1) corresponding to the posted speed N A = α, and to the real speed N R .
The end of the signal T o triggers a third monostable circuit 11 which produces a rectangular pulse signal T s (FIG. 5), of constant, short, duration. This is called the storage time. During the storage time T s , the voltage from capacitor 7 is compared with the C s voltage maintained at circuit 12, and transferred to 12 through a filter 12a so that the voltage maintained at 12 will quickly follow wide variations of the end-of-charge voltage from capacitor 7, but is not affected by small gaps due to irregularities of the period of T o . This voltage corresponding to the quantity q s to be injected as a function of the curve of statism C s is compared at a circuit 13 with the voltage corresponding to the polygonal contour representing the curve of full load C pc . Only the smaller of the two values is transmitted to a circuit 14 by comparator 13, and stored there.
Two rectangular signals D and δ, defined below, are delivered, for example by two circuit means 15 and 16. The rectangular signal D is applied to a delay circuit 17 having an inverter gate 18 at its input and a NAND gate 19 at its output. The signal D is also directly applied to an input of the NAND gate 19 whose output feeds a circuit branch 20 which in turn feeds one input of a bistable circuit 21, the other input of which is supplied by signal δ . The output signal of the bistable circuit 21, is the T auto (authorization) signal, (FIG. 6) and triggers the reading of storage circuit 14. The T auto signal also triggers the delivery by a pulse generator circuit 22 of a rectangular signal T inj whose duration is the duration of fuel injection.
To continue the description of the process in detail, at the end of T A , the current i(α, N R ), in which N R is the real speed of the motor, is applied to previously discharged capacitor 7 of FIG. 3 to charge it according to a linear law until the end of T o . This capacitor is charged during time (T o - T A ). The level of voltage reached, represents the quantity q s to be injected according to the curve of statism C s (FIG. 1) determined by the posted speed N A = α: ##SPC1##
in which C represents the capacitance of capacitor 7.
For the electronic circuity, the most complicated case is that in which the curves of statism are straight lines with constant inclination, or slope, that is to say that the partial derivative of q s with respect to N R d qs /dN R is independent of N R and N A , that is to say of α.
This can be done by
T A = A/αand i (α, N R ) - B . α . N R
We thus obtain, with T o = A/N R
q s = (B/C) α N R [(A/N R ) - (A/60 )]= (B/C) α N R . A (α - N R )/α N R
q s = (AB/C) (α - N R )
(δ q s /δ N R ) = - AB/C
In the event in which T A = K (constant)
i (α , N R ) = B . α . N R
We obtain straight curves of statism with steeper slope for larger α.
Finally, if T A - A/α and i (α, N R ) = D α,
the curves of statism are hyperbolic in form and the slope is greater for a higher posted speed α.
This shows it is possible to obtain a form of curve of statism adapted to the need of the motor, the three cases discussed being only examples among the possible embodiments without departing from the present invention.
An example of a circuit for the elements 5 to 13 of FIG. 3, is detailed in FIG. 7. Elements 9-10 comprise a potentiometer whose slider is grounded and has a total resistance R. For a given speed α, as determined by the setting of the slider, branch 9 has a resistance (1-α)R, and branch 10, a resistance αR.
The circuit for current generator 8 essentially comprises a three stage amplifier formed by transistors Q1, Q2, Q3. The point common to the collector of transistor Q1 and the base of transistor Q2 is connected to the branch 10 of the potentiometer. The collector of transistor Q3 is connected to capacitor C7 which forms the circuit 7 of FIG. 3.
The monostable circuit 5 (FIG. 3) is formed by transistors Q10 to Q15. The point common to the collector of transistor Q11 and the base of transistor Q12 are connected to the branch 9 of the potentiometer. The signal T o coming from means 1 is applied to the base of transistor Q13. The rectangular waveform signal T A is produced on the collector of Q15 from where it is applied to circuit 6.
The circuit 6 for producing the signal (T o - T A ), or Δ T, is formed by transistors Q16 and Q17. Signal T o is applied, by a diode, to the collector of transistor Q16 and signal T A is applied to its base. The signal (T o - T A ) is available on the collector of Q17 from where it is applied to the emitter of transistor Q2 of the current generator 8.
The monostable circuit 11 includes transistor Q18 whose collector is connected through a diode to the upper terminal of capacitor C7. A transistor Q19 has its collector connected to the emitter of transistor Q18. A capacitor C11 receives the signal T o on one electrode. The other electrode of C11 is connected to the base of transistor Q19 by a diode and also to the positive voltage supply (+) by a resistor. The base of Q18 receives the (T o - T A ) signal from the collector of Q17.
The rectangular waveform signal T s , defining the storage time, is produced by the trailing edge of T o and is available on the collector of transistor Q19. It is applied to the base of transistor Q5, and to the cathode of diode D12 of circuit 12 for transfer, filtering and storage of the voltage corresponding to the quantity of fuel to be injected, as defined by the statism curve. Circuit 12 comprises transistors Q4 to Q7, which are connected with a certain number of diodes as a non-linear filter, and a storage capacitor C12.
Comparator circuit 13 is formed by two symmetrical transistors Q8 and Q9, the emitters of which are connected. The base of transistor Q8 receives a voltage from capacitor 12 corresponding to the curve of statism. The base of transistor Q8 receives a voltage from element 4 corresponding to the curve of full load. The smaller of two voltages is transmitted from the point common of the two emitters to the storage circuit 14.
The functioning of the circuit of FIG. 7 is as follows. The base of transistor Q1 (circuit 8) receives a voltage ΔV = μN. with respect to the positive supply voltage. This voltage is proportional to the real speed (N R ) of the motor. Transistor Q1 produces a current proportional to this voltage through the collector load resistor 10, whose value is αR. On the base of transistor Q2 therefore there is applied a voltage proportional to the product αN R .
During the time interval ΔT - T o - T A , transistor Q3 furnishes to capacitor C7, a current
i = B . α . N R , where B is the current gain of the transistor Q3.
Circuit 5 delivers a rectangular waveform signal T A of duration inversely proportional to α, using branch 9 of the potentiometer means representative of posted speed, whose resistance is (1-α) R.
It should be noted that the maximum possible charge voltage of capacitor C7 represents a quantity of fuel to be injected that is larger than the maximum quantity of the curve of full load (FIG. 2). In this manner, if (T o - T A ) is very large, hence the posted speed N A very much higher than the real speed N R , then computation of the quantity of fuel to be injected according to the posted curve of statism gives a result larger than that obtained by computation of the full-load quantity. When this too occurs, in comparator 13 the smaller quantity of the two is retained.
At the end of time T o , and during the short and constant-duration pulse T s , the voltage of capacitor C7 is held and read.
By reason of irregularities in running of the motor, or of vibrations, for example, the duration T o can vary from reading to reading, without a change in the mean speed of the motor N R , that is to say without the need to change the quantity of fuel to be injected. The end-of-charge voltages of capacitor C7 therefore can vary from one reading of T o to the next. These variations must be eliminated but without affecting the response time in case of sudden change in N R . This is accomplished by the non-linear filter of the circuit 12 which is a low-pass filter with limit frequency depending on the amplitude of the input signal. During the storage time T s , capacitor C7 is connected through the non-linear filter 12 with the condenser storing the statism quantity C12.
The different signals T o , T A , Δ T and T s , are represented in FIG. 5.
Referring to FIG. 8, the monostable circuit 2 comprises two transistors Q20 and Q21, the base of Q21 being connected to the collector of Q20 by a capacitor C2 and a diode. The base of Q20 receives a signal T o from circuit means 1. The collector of transistor Q21 delivers a rectangular signal T v of constant duration.
The T v signal is transformed in the filter network 3. This network comprises, essentially, capacitor network 3. This netword comprises, essentially, capacitor C3, diode D3 and resistors R3 and R4. This network delivers a voltage proportional to the speed of rotation of the motor, comprising a substantial alternating component due to the irregularities mentioned above. This voltage is applied to the non-linear filter formed essentially by the two transistors Q22 and Q23. At the output of this filter, the signal is smoothed out by capacitor C4, and at point 23, there is produced a voltage proportional to the mean speed of the motor. Transistor Q24 provides temperature compensation and on its emitter there appears the voltage V = λ N which is proportional to the mean speed of the motor and is applied to circuits 4 and 17.
Transistors Q25 and Q26 generate a voltage ΔV = μN with respect to the positive feed, proportional to the motor mean speed, and corresponding to the voltage V = λN with respect to the circuit common potential point. This signal ΔV is applied to circuits 8 and 17. These two voltages V (λN) and ΔV (μN) are available continuously, and for a constant mean speed of the motor they are constant.
The voltage V = λN is applied to the function generator 4, formed by a number of current generators. The current generators include transistors Q27 to Q33 and adjusting potentiometers P1 to P6. One or more of the transistors Q27 to Q33 are connected with a respective potentiometer P1 to P6 as a series of current generators each of which is set to conduct at a predetermined level of the voltage V = λN and sends, to an adder-subtractor circuit Q30, the current from the current generator which corresponds to it. The current produced by each generator is proportional to the difference between the voltage V = λN and the threshold of the current generator considered. On the collector of transistor Q30, a voltage f(N) is available, corresponding to the full load curve (FIG. 2) which is applied to the base of Q9 of the comparator 13.
A function generator 4 of the type described can act as safety by deriving all the preceding currents and bringing the output level thereof to the level representing a zero quantity of fuel to be injected when the real speed of the motor exceeds the upper limit. Moreover, by means of other sets of current generators it is possible to be able to add or subtract currents as a function of outside parameters, such as, for example, air feed pressure, motor starting, etc. This will change the shape of the curve of FIG. 2.
The comparator 13, which receives the full load curve from generator 4 at Q9 and the statism voltage C s at Q8 from C12 (FIG. 7) transmits the smaller of the two voltages applied to it. The output of comparator 13 is the voltage V inj which corresponds to the quantity of fuel to be injected.
The V inj voltage is stored in circuit 14 (FIG. 9). In the absence of an authorization signal from circuit 21 (FIG. 3), transistor Q34 charges capacitor C14 at voltage V inj . When the authorization signal from circuit 21 appears, transistor Q34 blocks and capacitor C14 discharges through resistor R22.
Circuit 22 of FIG. 9 is a Schmidt trigger whose signal waveform has steep edges. The Schmidt circuit also has high input impedance. It is composed essentially of transistors Q35 to Q38. The emitters of transistors Q35 and Q36 are connected together. A diode D22, delivers the authorization signal from circuit 21 to Q35 and Q36. The appearance of the authorization signal through D22 makes transistors Q36 and Q37 conductive and blocks transistor Q38. With Q38 blocked the injection signal T inj appears on its collector. Transistor Q35 is blocked as long as the voltage at the upper electrode of capacitor C14 remains above a certain threshold level. When this voltage drops below the threshold level, transistor Q35 begins to conduct, which blocks transistors Q36 and Q37, and makes transistor Q38 conductive, thereby terminating the injection signal T inj .
The duration of T inj is therefore a function of the level of charge of capacitor C14 just before the start of the injection. The charge level of C14 is determined before the injection, the information has therefore been stored before use. The injection duration is determined by the discharge of capacitor C14 in circuit with resistor R22. This discharge is read by transistors Q35 to Q38.
The injection signal is, moreover, stopped no later than at the end of the authorization signal by the grounding through diode D22 of the emitter of transistor Q35.
The injection signal must be delivered to each injector at a time corresponding to a certain angular position of the camshaft. This time, which differs from one injector to another, is not fixed. Normally, it is a function of the speed of rotation of the motor, or of the quantity of fuel to be injected as well.
The circuit in FIG. 9 permits the control of the times of injection according to a certain law. These times, or injection offset angles, are designated by φ, which can vary between φ 0 and φ 1 . The law selected to define the offset angle is the following:
φ = φ 1 , for N R < Ni
φ = φ 0 + K 1 - K 2 . N R , for N R ≥ Ni
The curve representative of this function of level height φ 1 between zero and Ni, and a straight line with negative slope from Ni and to a value of the speed corresponding to φ = φ 0 .
As we saw before, the discharge of capacitor C14 begins when the authorization signal is high and blocks Q34. The authorization signal stops all injection when it is low. This low-level passage is, moreover, necessary in order to permit the recharge of C14.
The angular offset φ of the start of injection is therefore effected by an angular offset φof the start of authorization time.
The formation of the authorization signal is effected by a transformation of the two signds D and δ (see FIG. 6) which are defined in the following manner.
D is a rectangular pulse periodic signal whose total period is equal to that of the motor's cycle. Within this total period, the D pulse signal is reproduced as many times as there are injections to be made. The angular offset and the duration of each positive going pulse of the D signal are such that the start of each pulse corresponds to φ 0 and at the end to φ 1 . This is shown in FIG. 4.
This rectangular signal D is applied to the retard or delay, circuit 17 to produce a T AV signal defined by:
T AV = φ AV . (k/N R ), where
φ is the angular offset as defined above
φ 0 is the initial angle,
φ AV is the difference of these two angles,
T AV corresponds to a time duration,
N R is the angular speed and
k is a coefficient of homogeneity.
Circuit 17 furnishes a rectangular signal (T AV ) of the same total angular period as that of signal D, having as many negative going pulses as there are injections. Each rising front of a pulse D is matched by a descending front of a pulse T AV (FIG. 6).
The δ signals are positive going rectangular pulses of greater duration than the positive going pulse portion of the D signals. However, the δ signal overlaps the start and end of the D signal, as shown in FIG. 4.
The signals D and δ can be obtained directly by a means operating with the camshaft, for example, a mechanical or photoelectric switching arrangement. It is also possible to obtain them from other data. An important feature of the invention is the form and the offset of the signals D and δ.
A difficulty encountered in constructing a circuit to correct the injection time by a delay time, is in making the delay very large for very low speeds of rotation N R , and particularly when starting the motor.
One feature of the invention includes using, in addition to the angular marking of shaft position characterized by φ 0 , a second angular marking such that the start of an injection will never be started after the detection of this second mark, which therefore corresponds to φ 1 . Thi is the reason why the signal D has the form defined above. The start of injection must therefore lie at
(φ 0 + T AV N R /k) with
T AV = (k 1 - k 2 N R ) (k/N R = B/N R ) - A, but not later than at φ 1 .
FIG. 9 shows the retard circuit 17 formed by eight transistors Q39 to Q46 and a capacitor C17. Transistors Q39 and Q40 have their respective bases and their respective collectors connected together. The voltage ΔV = μN generated by circuit 3 (FIG. 8) is applied to the bases of Q39 and Q40. Transistor Q41 receives the voltage V = λN at its base from circuit 3. The collector of Q41 is connected to the base of transistor Q42 whose collector is connected to the collectors of transistors Q39 and Q40 at common point 24. Transistor Q43 receives, at its base, the signal D coming from the reversing, or inverter gate 18. Any suitable logic gate circuit can be used. The collector of Q43 is connected to a common point 24 by a potentiometer P. A capacitor 17 is connected between common point 24 and ground. Common point 24 is likewise connected to the base of transistor Q44.
The emitters of transistors Q44 and Q45 are connected to the same point of potential by a pair of diodes. The collector of transistor Q45 is connected to the base of transistor Q46 and the base of Q45 is connected by means of a resistor and a diode to the collector of Q46 at a point 25. At point 25 a signal is available, generated by circuit 17, and applied to the NAND gate 19 to produce signal T AV .
As long as signal D is low, transistor Q43 is conductive and traversed by a current i from transistors Q39 and Q40. The voltage at common point 24 is therefore, at the upper limit, V o = P = i, where P is the resistance value of the potentiometer.
When signal D becomes high, transistor Q43 blocks, and current i charges capacitor C17 linearly. When the voltage on C17 reaches a certain level fixed by the base potential of transistor Q45, the Schmidt trigger formed by transistors Q44 to Q46 fires and the potential of point 25 falls, indicating the end of the signal T AV .
Capacitor C17 is charged linearly by a current i = aN R proportional to V = λN since current i comes from transistors Q39 and Q40 whose emitters are connected to the supply voltage and whose bases receive the voltage ΔV = μN, Capacitor C17 is charged from a level V o = P . i . and to the threshold level V ref . of the trigger: ##SPC2##
T AV = C/i (V ref - V o ) = (C V ref - bN R /a N R )
T AV = [(C . V ref /a . 1/N R ) - b/a]
The potential on the collector of the transistor Q46 rises for a certain time after the end of the positive pulse of D, but this does not interfere in the circuit.
The signal at point 25, from circuit 17, and signal D are applied to NAND gate 19. At the output of NAND 19 a signal appears which is high up to φ 0 and which is low for a time T AV , but φ goes back up to a high value before φ 1 .
Transistors Q41 and Q42 form a trigger which trips, for a certain value Ni of N, to control, or shape, the current furnished by Q39 and Q40 to charge C17. Q41 and Q42 insure that below speed Ni, the charge of C17 will be slow, slower than normal. In the case when Q41 and Q42 operate the end of T AV will no longer be controlled by the tripping of the trigger of Q44 to Q46, but by the end of signal D, which changes the state of Q43.
The signal T AV is differentiated by a circuit 20 to deliver a positive pulse, or spike, when it becomes high. This is shown by line δ/δt in FIG. 6. The δ/δt pulse therefore appears each time an injection is to begin.
The signal δ which is rectangular and periodic like D. Each of the positive pulses of δ frames a positive pulse of D, and the injection, if any, must necessarily take place within the positive pulse of D. This means that δ becomes high no later than at the same time as D, and becomes low in the position characterizng the extreme limit of injection with maximum delay, i.e. the maximum delay plus the maximum injection duration tolerated.
The start-of-injection pulse δ/δt triggers a bistable circuit 21 which returns to its initial state at the end of the signal δ. The bistable circuit 21 produces the T auto pulse.
This bistable circuit 21 can be embodied simply with two NAND gates, the first receiving the signals from 16 and 20 (FIG. 3), the second receiving the signal furnished by the first, a loop being closed between the output of the second gate and the input of the first coming from differentiating circuit 20. The output signal of circuit 21 represents the authorization time T auto applied to circuits 14 and 22 (FIG. 3).
The series of signals D, δ signal at point 25, T AV , differentiated signal δ/δt and T auto , is represented in FIG. 6.
One of the principal advantages of the device according to the invention is that it permits regularizing the injections by taking a mean value for the speed of rotation of the motor, and not momentary value, which is essentially variable by jolts.
The problem to be solved by a computer for a fuel injection system is essentially that of electronic determination of the quantity of fuel to be injected as a function of the speed of the motor, the posted speed and, optionally, other parameters. The term posted speed is used to designate the ordered speed, imposed for example by the use of the motor with the aid of a control organ, such as the accelerator. That is, for every position of the accelerator or other control organ there is a theoretical speed for the engine. It is known to electrically compute the quantity of fuel to be injected as a function of the speed of the motor and the posted speed. The signal produced by he computation is generally a voltage or a current, which is continually available. This type of computation is used with conventional fuel injection pumps to position the pump regulating mechanism.
It is also known, in the case of electronic fuel injection systems, about computing, whenever an injection is to take place, the duration of this injection. In other words, each time an injection is started or triggered, the computation of its duration is started by a measurement of all variables. The injection is stopped at the end of this duration.
It is also known to elaborate or compute and produce, before the injection, a signal representative of the duration of the injection, storing it in a suitable circuit, and "reading" it in the course of the injection so that the actual duration of the injection will correspond to that produced and stored.
The object of the invention is to provide a fuel injection duration computer which computes the injection duration information before the injection, and stores this information in a circuit for reading during the injection. The computer operates so that the computation is not integrally effected during the interval separating the injections, and the measurement of certain variables and certain preliminary calculations are effected on a continuous basis so that the intermediate results are available at all times.
The advantage of such a design is that it is thereby possible to make the measurements with greater precision and to process the data measured in such a way as to insure both precision and a sufficiently short response time. This is done while retaining the advantages of repeated computation, which are, primarily, lower sensitivity to electrical parasites, or interference signal, better stability, and the presence of periods when injection is authorized, the latter insuring security against eroneous injections.
The object of the invention is an electronic computer for a system of fuel injection for combustion engines of the type in which the electrical data corresponding to the quantity of fuel to be injected is determined before the injection and stored in a circuit for reading during the injection. The computer includes a first means for delivering pulses at a frequency proportional to the speed of the motor and a first circuit for transforming these pulses into an electrical voltage proportional to the speed of the motor. A second means is also provided to introduce a signal corresponding to the posted speed of the motor. A second circuit delivers after each pulse from the first means, a voltage representative of the quantity of fuel to be injected according to the "statism" curves. The second circuit produces its voltage by transforming the signals from the first and second means and, optionally, the voltage which is proportional to the speed of the motor. A third circuit also is provided which transforms the voltage proportional to the speed of the motor into a voltage representing the quantity of fuel to be injected according to a curve corresponding to the full load operation of the motor. Also provided is a fourth circuit which transmits the smallest of the two voltages representative of the quantities of fuel to be injected according to the curves of the statism (produced by the second circuit) and according to the curve of full load (produced by the third circuit).
Other characteristics and objects of the invention will appear in the description which follows, made in reference to the attached drawings wherein:
FIG. 1 is a group of curves representative of the quantity of fuel to be injected as a function of the speed for a compression-ignition motor;
FIG. 2 is a polygonal contour approaching the curve of full load C pc of FIG. 1, and capable of being computed and produced by a function generator;
FIG. 3 is a schematic block diagram of the system according to the invention, in which the continuous data is represented by dotted lines, the pulsed data by broken lines, and triggering signals by solid lines;
FIG. 4 is a diagram representing angular offsets of signals D and δ (delta), used for the production of certain electrical signals;
FIG. 5 is a timing diagram of the signals T o , T A , T S computed and produced by elements 1, 5 and 11 of the circuit of FIG. 3;
FIG. 6 is a timing diagram of distribution of signals D to T auto , computed and produced by elements 15 to 21 in FIG. 3;
FIG. 7 is a detailed diagram of elements 5 to 13 of FIG. 3, used for the computation and production of the voltage corresponding to the curves of statism;
FIG. 8 is a detailed diagram of elements 2 to 4 of FIG. 3, used for the computation and production of the voltage corresponding to the curve of full load; and
FIG. 9 is a detailed diagram of elements 17, 14 and 22 of FIG. 3, used for the computation and production of the injection duration.
In combustion motors, for example with fuel injection, the quantity q of fuel to be injected can be represented as a function of the speed of rotation N of the motor, by a group of curves such as that in FIG 1. In this figure, curve C pc represents the curve of full load for the motor, which can be exceeded only rarely. This complex curve can generally be approximated by a polygonal contour such as shown in FIG. 2.
The polygon of FIG. 2, corresponding to the full load curve, is varied or deformed as a function of the parameters of the motor (for example, supercharging with air at variable pressure), or for particular operating conditions (for example, starting). The elaboration, or generation, of the full load polygon of FIG. 2 is accomplished by an electronic function generator of any suitable known type.
In FIG. 1, the curves C s represent the curves of statism defined on the one hand by the posted speed of the motor N Ai , which is dictated by the user, and on the other hand by the function (dg/dN R) N Ai . The latter function depends, for a given curve, on the posted speed N Ai and on the real speed N R of the motor.
For a posted value N Ai of the speed, shown moving along one of the curves C s , and a real value of speed N R of the motor, shown by the dotten line, two values of the quantity of fuel to be injected can be determined. One is where the dotted line of N R intersects the curve of full load, at q pc and the other at point q s where it intersects the C s curve of statism N Ai .
The quantity of fuel to be actually injected is the smaller of two quantities q s and q pc . Naturally, if the real speed N R of the motor is higher than the posted speed N Ai , the quantity of fuel to be injected is zero.
In order to define the quantity of fuel to be injected, the real speed of the motor N R is measured as is the spread between the real and posted speeds (N R - N Ai ).
Depending upon the place at which the real speed is measured and the manner in which it is measured, the signal obtained contains an alternative or unwanted component due to an irregular movement of the motor component whose movement is monitored. An attempt can be made to eliminate this alternative component. However, this affects the response time of the measurement of the spread of the speeds.
Referring to FIG. 3, there is shown at 1, a fixed or static, means for delivering rectangular waveform signals of duration T o . The intervals T o of time corresponds to the time of passage before means 1, of a certain sector, of known dimensions, of a piece rotating synchronously with the motor, for example a sector on the fly wheel. This can be accomplished by a suitable electrical circuit or electromechanical switching means. It can also correspond to the interval of time separating the successive passages between two means, spaced by a known angular sector, of a reference mark rotating synchronously with the motor. This also can be done by suitable switching means as a photo-electric circuit. The interval of time T o is linked to the speed of rotation N R of the motor and to the angle A of opening of the sector considered. We have a relation of the type
T o = A/N R
the width (duration) of the rectangular signal T o being inversely proportional, and its frequency directly proportional to the speed of the motor. The T o signal is shown on the first line of FIG. 5.
The leading edge corresponding to the start of T o triggers a first monostable circuit 2 which delivers a rectangular waveform signal T v of constant duration. Since the speed of rotation of the motor is not always regular, but is subject to change, the period of signals T o , and the rate of production of signals T v is not exactly constant. The output signals T v of monostable circuit 2 is applied to a filter circuit 3 which transforms these signals into a voltage proportional to the speed N R . This voltage is also affected by a substantial spurious, or unwanted component. Filter 3 is therefore a non-linear filter which delivers a voltage V = λN, proportional to the mean speed of the motor. The voltage V is applied to a function generator 4 which delivers a voltage corresponding to the full load reading, and more exactly to the polygonal contour of FIG. 2.
The start of a T o signal also triggers a second monostable circuit 5 which produces a signal to determine a duration T A . The signal T A is a function of the value of the electrical signal representing the posted speed N A , defined below by α. The T A signal is shown on the second line of FIG. 5. In the interval of time (T o - T A ), as controlled by a circuit 6, and following signal T A , a capacitor 7 is charged linearly by a current generator 8 with a current i (α, N). The posted speed α is produced by circuits 9-10 furnishing two pieces of supplementary data: one, α to current generator 8, the other, (1- α 0), to monostable circuit 5. The current generator also receives the data of the duration (T o - T A ), or ΔT, and of the real speed of the motor V = λ N. The time ΔT is shown on the third line of FIG. 5. The level of voltage reached by the capacitor 7 represents the quantity of fuel q s to be injected according to the curve of statism C s (FIG. 1) corresponding to the posted speed N A = α, and to the real speed N R .
The end of the signal T o triggers a third monostable circuit 11 which produces a rectangular pulse signal T s (FIG. 5), of constant, short, duration. This is called the storage time. During the storage time T s , the voltage from capacitor 7 is compared with the C s voltage maintained at circuit 12, and transferred to 12 through a filter 12a so that the voltage maintained at 12 will quickly follow wide variations of the end-of-charge voltage from capacitor 7, but is not affected by small gaps due to irregularities of the period of T o . This voltage corresponding to the quantity q s to be injected as a function of the curve of statism C s is compared at a circuit 13 with the voltage corresponding to the polygonal contour representing the curve of full load C pc . Only the smaller of the two values is transmitted to a circuit 14 by comparator 13, and stored there.
Two rectangular signals D and δ, defined below, are delivered, for example by two circuit means 15 and 16. The rectangular signal D is applied to a delay circuit 17 having an inverter gate 18 at its input and a NAND gate 19 at its output. The signal D is also directly applied to an input of the NAND gate 19 whose output feeds a circuit branch 20 which in turn feeds one input of a bistable circuit 21, the other input of which is supplied by signal δ . The output signal of the bistable circuit 21, is the T auto (authorization) signal, (FIG. 6) and triggers the reading of storage circuit 14. The T auto signal also triggers the delivery by a pulse generator circuit 22 of a rectangular signal T inj whose duration is the duration of fuel injection.
To continue the description of the process in detail, at the end of T A , the current i(α, N R ), in which N R is the real speed of the motor, is applied to previously discharged capacitor 7 of FIG. 3 to charge it according to a linear law until the end of T o . This capacitor is charged during time (T o - T A ). The level of voltage reached, represents the quantity q s to be injected according to the curve of statism C s (FIG. 1) determined by the posted speed N A = α: ##SPC1##
in which C represents the capacitance of capacitor 7.
For the electronic circuity, the most complicated case is that in which the curves of statism are straight lines with constant inclination, or slope, that is to say that the partial derivative of q s with respect to N R d qs /dN R is independent of N R and N A , that is to say of α.
This can be done by
T A = A/αand i (α, N R ) - B . α . N R
We thus obtain, with T o = A/N R
q s = (B/C) α N R [(A/N R ) - (A/60 )]= (B/C) α N R . A (α - N R )/α N R
q s = (AB/C) (α - N R )
(δ q s /δ N R ) = - AB/C
In the event in which T A = K (constant)
i (α , N R ) = B . α . N R
We obtain straight curves of statism with steeper slope for larger α.
Finally, if T A - A/α and i (α, N R ) = D α,
the curves of statism are hyperbolic in form and the slope is greater for a higher posted speed α.
This shows it is possible to obtain a form of curve of statism adapted to the need of the motor, the three cases discussed being only examples among the possible embodiments without departing from the present invention.
An example of a circuit for the elements 5 to 13 of FIG. 3, is detailed in FIG. 7. Elements 9-10 comprise a potentiometer whose slider is grounded and has a total resistance R. For a given speed α, as determined by the setting of the slider, branch 9 has a resistance (1-α)R, and branch 10, a resistance αR.
The circuit for current generator 8 essentially comprises a three stage amplifier formed by transistors Q1, Q2, Q3. The point common to the collector of transistor Q1 and the base of transistor Q2 is connected to the branch 10 of the potentiometer. The collector of transistor Q3 is connected to capacitor C7 which forms the circuit 7 of FIG. 3.
The monostable circuit 5 (FIG. 3) is formed by transistors Q10 to Q15. The point common to the collector of transistor Q11 and the base of transistor Q12 are connected to the branch 9 of the potentiometer. The signal T o coming from means 1 is applied to the base of transistor Q13. The rectangular waveform signal T A is produced on the collector of Q15 from where it is applied to circuit 6.
The circuit 6 for producing the signal (T o - T A ), or Δ T, is formed by transistors Q16 and Q17. Signal T o is applied, by a diode, to the collector of transistor Q16 and signal T A is applied to its base. The signal (T o - T A ) is available on the collector of Q17 from where it is applied to the emitter of transistor Q2 of the current generator 8.
The monostable circuit 11 includes transistor Q18 whose collector is connected through a diode to the upper terminal of capacitor C7. A transistor Q19 has its collector connected to the emitter of transistor Q18. A capacitor C11 receives the signal T o on one electrode. The other electrode of C11 is connected to the base of transistor Q19 by a diode and also to the positive voltage supply (+) by a resistor. The base of Q18 receives the (T o - T A ) signal from the collector of Q17.
The rectangular waveform signal T s , defining the storage time, is produced by the trailing edge of T o and is available on the collector of transistor Q19. It is applied to the base of transistor Q5, and to the cathode of diode D12 of circuit 12 for transfer, filtering and storage of the voltage corresponding to the quantity of fuel to be injected, as defined by the statism curve. Circuit 12 comprises transistors Q4 to Q7, which are connected with a certain number of diodes as a non-linear filter, and a storage capacitor C12.
Comparator circuit 13 is formed by two symmetrical transistors Q8 and Q9, the emitters of which are connected. The base of transistor Q8 receives a voltage from capacitor 12 corresponding to the curve of statism. The base of transistor Q8 receives a voltage from element 4 corresponding to the curve of full load. The smaller of two voltages is transmitted from the point common of the two emitters to the storage circuit 14.
The functioning of the circuit of FIG. 7 is as follows. The base of transistor Q1 (circuit 8) receives a voltage ΔV = μN. with respect to the positive supply voltage. This voltage is proportional to the real speed (N R ) of the motor. Transistor Q1 produces a current proportional to this voltage through the collector load resistor 10, whose value is αR. On the base of transistor Q2 therefore there is applied a voltage proportional to the product αN R .
During the time interval ΔT - T o - T A , transistor Q3 furnishes to capacitor C7, a current
i = B . α . N R , where B is the current gain of the transistor Q3.
Circuit 5 delivers a rectangular waveform signal T A of duration inversely proportional to α, using branch 9 of the potentiometer means representative of posted speed, whose resistance is (1-α) R.
It should be noted that the maximum possible charge voltage of capacitor C7 represents a quantity of fuel to be injected that is larger than the maximum quantity of the curve of full load (FIG. 2). In this manner, if (T o - T A ) is very large, hence the posted speed N A very much higher than the real speed N R , then computation of the quantity of fuel to be injected according to the posted curve of statism gives a result larger than that obtained by computation of the full-load quantity. When this too occurs, in comparator 13 the smaller quantity of the two is retained.
At the end of time T o , and during the short and constant-duration pulse T s , the voltage of capacitor C7 is held and read.
By reason of irregularities in running of the motor, or of vibrations, for example, the duration T o can vary from reading to reading, without a change in the mean speed of the motor N R , that is to say without the need to change the quantity of fuel to be injected. The end-of-charge voltages of capacitor C7 therefore can vary from one reading of T o to the next. These variations must be eliminated but without affecting the response time in case of sudden change in N R . This is accomplished by the non-linear filter of the circuit 12 which is a low-pass filter with limit frequency depending on the amplitude of the input signal. During the storage time T s , capacitor C7 is connected through the non-linear filter 12 with the condenser storing the statism quantity C12.
The different signals T o , T A , Δ T and T s , are represented in FIG. 5.
Referring to FIG. 8, the monostable circuit 2 comprises two transistors Q20 and Q21, the base of Q21 being connected to the collector of Q20 by a capacitor C2 and a diode. The base of Q20 receives a signal T o from circuit means 1. The collector of transistor Q21 delivers a rectangular signal T v of constant duration.
The T v signal is transformed in the filter network 3. This network comprises, essentially, capacitor network 3. This netword comprises, essentially, capacitor C3, diode D3 and resistors R3 and R4. This network delivers a voltage proportional to the speed of rotation of the motor, comprising a substantial alternating component due to the irregularities mentioned above. This voltage is applied to the non-linear filter formed essentially by the two transistors Q22 and Q23. At the output of this filter, the signal is smoothed out by capacitor C4, and at point 23, there is produced a voltage proportional to the mean speed of the motor. Transistor Q24 provides temperature compensation and on its emitter there appears the voltage V = λ N which is proportional to the mean speed of the motor and is applied to circuits 4 and 17.
Transistors Q25 and Q26 generate a voltage ΔV = μN with respect to the positive feed, proportional to the motor mean speed, and corresponding to the voltage V = λN with respect to the circuit common potential point. This signal ΔV is applied to circuits 8 and 17. These two voltages V (λN) and ΔV (μN) are available continuously, and for a constant mean speed of the motor they are constant.
The voltage V = λN is applied to the function generator 4, formed by a number of current generators. The current generators include transistors Q27 to Q33 and adjusting potentiometers P1 to P6. One or more of the transistors Q27 to Q33 are connected with a respective potentiometer P1 to P6 as a series of current generators each of which is set to conduct at a predetermined level of the voltage V = λN and sends, to an adder-subtractor circuit Q30, the current from the current generator which corresponds to it. The current produced by each generator is proportional to the difference between the voltage V = λN and the threshold of the current generator considered. On the collector of transistor Q30, a voltage f(N) is available, corresponding to the full load curve (FIG. 2) which is applied to the base of Q9 of the comparator 13.
A function generator 4 of the type described can act as safety by deriving all the preceding currents and bringing the output level thereof to the level representing a zero quantity of fuel to be injected when the real speed of the motor exceeds the upper limit. Moreover, by means of other sets of current generators it is possible to be able to add or subtract currents as a function of outside parameters, such as, for example, air feed pressure, motor starting, etc. This will change the shape of the curve of FIG. 2.
The comparator 13, which receives the full load curve from generator 4 at Q9 and the statism voltage C s at Q8 from C12 (FIG. 7) transmits the smaller of the two voltages applied to it. The output of comparator 13 is the voltage V inj which corresponds to the quantity of fuel to be injected.
The V inj voltage is stored in circuit 14 (FIG. 9). In the absence of an authorization signal from circuit 21 (FIG. 3), transistor Q34 charges capacitor C14 at voltage V inj . When the authorization signal from circuit 21 appears, transistor Q34 blocks and capacitor C14 discharges through resistor R22.
Circuit 22 of FIG. 9 is a Schmidt trigger whose signal waveform has steep edges. The Schmidt circuit also has high input impedance. It is composed essentially of transistors Q35 to Q38. The emitters of transistors Q35 and Q36 are connected together. A diode D22, delivers the authorization signal from circuit 21 to Q35 and Q36. The appearance of the authorization signal through D22 makes transistors Q36 and Q37 conductive and blocks transistor Q38. With Q38 blocked the injection signal T inj appears on its collector. Transistor Q35 is blocked as long as the voltage at the upper electrode of capacitor C14 remains above a certain threshold level. When this voltage drops below the threshold level, transistor Q35 begins to conduct, which blocks transistors Q36 and Q37, and makes transistor Q38 conductive, thereby terminating the injection signal T inj .
The duration of T inj is therefore a function of the level of charge of capacitor C14 just before the start of the injection. The charge level of C14 is determined before the injection, the information has therefore been stored before use. The injection duration is determined by the discharge of capacitor C14 in circuit with resistor R22. This discharge is read by transistors Q35 to Q38.
The injection signal is, moreover, stopped no later than at the end of the authorization signal by the grounding through diode D22 of the emitter of transistor Q35.
The injection signal must be delivered to each injector at a time corresponding to a certain angular position of the camshaft. This time, which differs from one injector to another, is not fixed. Normally, it is a function of the speed of rotation of the motor, or of the quantity of fuel to be injected as well.
The circuit in FIG. 9 permits the control of the times of injection according to a certain law. These times, or injection offset angles, are designated by φ, which can vary between φ 0 and φ 1 . The law selected to define the offset angle is the following:
φ = φ 1 , for N R < Ni
φ = φ 0 + K 1 - K 2 . N R , for N R ≥ Ni
The curve representative of this function of level height φ 1 between zero and Ni, and a straight line with negative slope from Ni and to a value of the speed corresponding to φ = φ 0 .
As we saw before, the discharge of capacitor C14 begins when the authorization signal is high and blocks Q34. The authorization signal stops all injection when it is low. This low-level passage is, moreover, necessary in order to permit the recharge of C14.
The angular offset φ of the start of injection is therefore effected by an angular offset φof the start of authorization time.
The formation of the authorization signal is effected by a transformation of the two signds D and δ (see FIG. 6) which are defined in the following manner.
D is a rectangular pulse periodic signal whose total period is equal to that of the motor's cycle. Within this total period, the D pulse signal is reproduced as many times as there are injections to be made. The angular offset and the duration of each positive going pulse of the D signal are such that the start of each pulse corresponds to φ 0 and at the end to φ 1 . This is shown in FIG. 4.
This rectangular signal D is applied to the retard or delay, circuit 17 to produce a T AV signal defined by:
T AV = φ AV . (k/N R ), where
φ is the angular offset as defined above
φ 0 is the initial angle,
φ AV is the difference of these two angles,
T AV corresponds to a time duration,
N R is the angular speed and
k is a coefficient of homogeneity.
Circuit 17 furnishes a rectangular signal (T AV ) of the same total angular period as that of signal D, having as many negative going pulses as there are injections. Each rising front of a pulse D is matched by a descending front of a pulse T AV (FIG. 6).
The δ signals are positive going rectangular pulses of greater duration than the positive going pulse portion of the D signals. However, the δ signal overlaps the start and end of the D signal, as shown in FIG. 4.
The signals D and δ can be obtained directly by a means operating with the camshaft, for example, a mechanical or photoelectric switching arrangement. It is also possible to obtain them from other data. An important feature of the invention is the form and the offset of the signals D and δ.
A difficulty encountered in constructing a circuit to correct the injection time by a delay time, is in making the delay very large for very low speeds of rotation N R , and particularly when starting the motor.
One feature of the invention includes using, in addition to the angular marking of shaft position characterized by φ 0 , a second angular marking such that the start of an injection will never be started after the detection of this second mark, which therefore corresponds to φ 1 . Thi is the reason why the signal D has the form defined above. The start of injection must therefore lie at
(φ 0 + T AV N R /k) with
T AV = (k 1 - k 2 N R ) (k/N R = B/N R ) - A, but not later than at φ 1 .
FIG. 9 shows the retard circuit 17 formed by eight transistors Q39 to Q46 and a capacitor C17. Transistors Q39 and Q40 have their respective bases and their respective collectors connected together. The voltage ΔV = μN generated by circuit 3 (FIG. 8) is applied to the bases of Q39 and Q40. Transistor Q41 receives the voltage V = λN at its base from circuit 3. The collector of Q41 is connected to the base of transistor Q42 whose collector is connected to the collectors of transistors Q39 and Q40 at common point 24. Transistor Q43 receives, at its base, the signal D coming from the reversing, or inverter gate 18. Any suitable logic gate circuit can be used. The collector of Q43 is connected to a common point 24 by a potentiometer P. A capacitor 17 is connected between common point 24 and ground. Common point 24 is likewise connected to the base of transistor Q44.
The emitters of transistors Q44 and Q45 are connected to the same point of potential by a pair of diodes. The collector of transistor Q45 is connected to the base of transistor Q46 and the base of Q45 is connected by means of a resistor and a diode to the collector of Q46 at a point 25. At point 25 a signal is available, generated by circuit 17, and applied to the NAND gate 19 to produce signal T AV .
As long as signal D is low, transistor Q43 is conductive and traversed by a current i from transistors Q39 and Q40. The voltage at common point 24 is therefore, at the upper limit, V o = P = i, where P is the resistance value of the potentiometer.
When signal D becomes high, transistor Q43 blocks, and current i charges capacitor C17 linearly. When the voltage on C17 reaches a certain level fixed by the base potential of transistor Q45, the Schmidt trigger formed by transistors Q44 to Q46 fires and the potential of point 25 falls, indicating the end of the signal T AV .
Capacitor C17 is charged linearly by a current i = aN R proportional to V = λN since current i comes from transistors Q39 and Q40 whose emitters are connected to the supply voltage and whose bases receive the voltage ΔV = μN, Capacitor C17 is charged from a level V o = P . i . and to the threshold level V ref . of the trigger: ##SPC2##
T AV = C/i (V ref - V o ) = (C V ref - bN R /a N R )
T AV = [(C . V ref /a . 1/N R ) - b/a]
The potential on the collector of the transistor Q46 rises for a certain time after the end of the positive pulse of D, but this does not interfere in the circuit.
The signal at point 25, from circuit 17, and signal D are applied to NAND gate 19. At the output of NAND 19 a signal appears which is high up to φ 0 and which is low for a time T AV , but φ goes back up to a high value before φ 1 .
Transistors Q41 and Q42 form a trigger which trips, for a certain value Ni of N, to control, or shape, the current furnished by Q39 and Q40 to charge C17. Q41 and Q42 insure that below speed Ni, the charge of C17 will be slow, slower than normal. In the case when Q41 and Q42 operate the end of T AV will no longer be controlled by the tripping of the trigger of Q44 to Q46, but by the end of signal D, which changes the state of Q43.
The signal T AV is differentiated by a circuit 20 to deliver a positive pulse, or spike, when it becomes high. This is shown by line δ/δt in FIG. 6. The δ/δt pulse therefore appears each time an injection is to begin.
The signal δ which is rectangular and periodic like D. Each of the positive pulses of δ frames a positive pulse of D, and the injection, if any, must necessarily take place within the positive pulse of D. This means that δ becomes high no later than at the same time as D, and becomes low in the position characterizng the extreme limit of injection with maximum delay, i.e. the maximum delay plus the maximum injection duration tolerated.
The start-of-injection pulse δ/δt triggers a bistable circuit 21 which returns to its initial state at the end of the signal δ. The bistable circuit 21 produces the T auto pulse.
This bistable circuit 21 can be embodied simply with two NAND gates, the first receiving the signals from 16 and 20 (FIG. 3), the second receiving the signal furnished by the first, a loop being closed between the output of the second gate and the input of the first coming from differentiating circuit 20. The output signal of circuit 21 represents the authorization time T auto applied to circuits 14 and 22 (FIG. 3).
The series of signals D, δ signal at point 25, T AV , differentiated signal δ/δt and T auto , is represented in FIG. 6.
One of the principal advantages of the device according to the invention is that it permits regularizing the injections by taking a mean value for the speed of rotation of the motor, and not momentary value, which is essentially variable by jolts.