EP0112673A1

EP0112673A1 - Electronic apparatus for controlling the supply of fuel to an internal combustion engine

Info

Publication number: EP0112673A1
Application number: EP83307471A
Authority: EP
Inventors: Brian Colin Pagdin; David John Marriage; Christopher John Feeney
Original assignee: Solex UK Ltd
Current assignee: Solex UK Ltd
Priority date: 1982-12-13
Filing date: 1983-12-08
Publication date: 1984-07-04
Also published as: JPS59165831A

Abstract

An electronic control unit is used to control the opening period of a single point fuel injector by means of the generation of a fuel metering electrical pulse. The pulse length is determined by the sensing of inter alia engine loading (e. g. air flow rate or absolute manifold pressure) and engine speed. The engine loading signal is sampled at fixed intervals and average over the samples taken between engine firings. The number of samples used in the averaging procedure reduces with increasing engine speed.

An acceleration enrichment procedure is also conducted by the control unit comprising a proportional increase in pulse width for a number of cycles followed by a smaller increase over successive cycles.

Description

This invention relates to electronic apparatus for controlling the supply of fuel to an internal combustion engine.
Electronic apparatus in accordance with the invention may be used to control internal combustion engines employing pressure differences to draw fucel into the cylinders via a carburetter or to those in which the fuel is supplied through one or more injection valves. In particular the invention relates to internal combustion engines in which fuel is supplied to the engine in discrete charges; the quantity of fuel in each charge being controlled by an electrical signal from a control unit to which a fuel valve is responsive and the frequency of the electrical signal being determined by engine speed. The.value of the electrical signal that determines the supplied quantity of fuel, is derived from the sensed values of certain engine parameters chosen to optimise the air/fuel mixture for differing operating conditions. Principal amongst these parameters are engine speed and a parameter that is some function of engine loading.
The air mass flow rate to the engine is one such function of ergine loading, and it can be sensed, for example, by sensing the displacement of a flap meter in the inlet manifold. Other parameters that are functions of engine loading are absolute manifold pressure, or in the case of a carburetter the amount of lift of the air valve. In known systems a transducer converts the values of the sensed parameter into an electrical signal which is used to interrogate a matrix in an electronic data store to read out a value for the fuel metering control signal. In the absence of mechanical damping or electrical smoothing, the electrical signal representing the function of engine loading such as mass flow rate at the inlet manifold is not constant over an engine cycle, but varies in a generally sinusoidal manner. The fuel metering control signal is not required to be modified by this variation and hence in known control systems the variation is removed by increased mechanical damping where a mechanical sensor is used or by RC filtering of the electrical signal. A major disadvantage of such damping or filtering is the consequent degradation of system response time. Hence, in a transient state changes in this signal will lag behind the changes in the actual values of the engine parameter.
A further disadvantage of such smoothing of the signal is the reduction in sensitivity to non-cyclic changes in for example the mass flow rate or manifold pressure.
In U.S. 4197767 methods of averaging a torque signal are disclosed for controlling an I/C engine. These however rely on running averages with & new value being obtained at the cylinder firing rate. This new value is however based on a single new value averaged with a number of values obtained previously. Thus variation within an engine cycle is not discriminated, and response time is likely to be relatively slow.
It is an object of the present invention to provide an improved method of deriving an electrical signal representing a function of engine loading by electronic processing of the signal from the relevant transducer.
According to the present invention in a first aspect there is provided electronic apparatus for controlling the supply of fuel to an internal combustion engine, the electronic apparatus being adapted to receive, in use, a first electrical signal representing a function of engine loading and a second electrical signal having, or representing, a periodicity derived from the cylinder firings of the engine, the electronic apparatus producing an output signal which, in use, is connected to a fuel metering pulse, said output signal having a value selected or derived by said electronic apparatus on the basis of at least the value of a signal representing an averaged value of the function of engine loading, characterised in that the electronic apparatus includes signal processing means operative to sample said first electrical signal at fixed intervals over a period fixed to the period of said second signal, and further operative to derive a signal representing an averaged value of said samples. The average value may be derived from said samples by a numerical integration method determined by a computer program controlling said signal processing means. The numerical integration method may be based, for example, on Simpson's Rule or a Trapezoidal Rule.
The electronic apparatus may include electronic counting means operative to count the number of sample values of sais first electrical signal during one or more periods of said second electrical signal, and hence derive a signal representing engine-speed, the value of said output signal from the electronic apparatus being selected or derived also on the basis of said signal representing engine speed.
The signal processing means may be further operative to compare sequential ones of said signals representing averaged values of the function of engine loading and to initiate an additional increase in the value of said output signal when said sequential signals are increasing in value above a predetermined amount or rate.
Use of electronic control apparatus enables a more sophisticated stragegy of fuel enrichment in response to acceleration to be carried out than with mechanically-controlled devices such as accelerator pumps. The Applicants have discovered that the addition of fuel in response to an indication of the command to accelerate achieves best results when the additional fuel is provided other than solely by a constant amou t over a number of engine cycles. The discovery may be applied in systems other than those using the sampling techniques as aforesaid. For example, it may also be applied to prior art systems in which intended acceleration is indicated by the rate of change of throttle angle.
Therefore, according to the invention in a second aspect there is provided electronic control apparatus for controlling the supply of fuel to an internal combustion engine including an input connected in use to receive a signal indicating intended or actual engine acceleration, the electronic control apparatus having an output at which, in use, a signal controlling a fuel metering pulse is produced, characterised in that the control apparatus is responsive to said signal indicating acceleration to increase the value of said output signal and hence the amount of fuel delivered to the engine, by a predetermined constant amount over a predetermined range of engine cycles and by a reduced amount over a subsequent range of engine cycles.
Embodiments of the invention will now be described, by way of example only, with reference to the accompanying drawings. In the drawings:-

Figure 1 is a schematic diagram of relevant mechanical parts of a single point fuel injection system to which the present invention may be applied.
Figure 2 is a block schematic diagram of an electronic control unit;
Figure 3 is a graph showing the variation with time of mass flow at the inlet manifold;
Figure 4 is a flow diagram representing the sequence of operations carried out by a microprocessor in the electronic control unit of Figure 2.
Figure 5 is a graphical illustration of fuel enrichment strategy; and
Figure 6 is a flow diagram representing an alternative sequence of operation carried out by a microprocessor in an electronic control unit.

In one example the invention is embodied in a single point fuel injection system applied to a four cylinder petrol engine. Figure 1 illustrates the relevant mechanical components. A single electrically-operated ball valve injector 1 is mounted to direct fuel into the inlet manifold 2 of the engine. The injection nozzle of the valve 1 is angled towards the downstream direction of air flow in the manifold. A throttle valve 3 is pivoted in the intake manifold chamber downstream of the injector 1 and in which position the injection nozzle is directed generally in the direction of the central pivotal axis of the throttle valve 3. ;
An output pipe 4 is connected from an anti-surge zone in a fuel tank 5 and feeds a fuel pump 6 via a filter unit 7. The fuel pump delivers fuel via a pipe 8 to a pressure regulator 9 close to the injector 1. Surplus fuel is returned to another portion of the fuel tank from the regulator 9 through a return pipe 10.
The quantity of fuel injected into the manifold by the injector 1 is determined by the duration of an electrical pulse produced by a solenoid drive circuit operating to lift the ball valve in the injector. The drive pulse comprises an initial ramp portion typically reaching 4 amps in 1 millisecond to move the ball of the valve, followed, after an inductive decay, by a 1 amp holding current for the appropriate period. The width of a signal pulse, developed by microprocessor-based control unit 20, determines the drive pulse duration. The signal pulse width is computed on the basis of the sensed values of a number of engine operating parameters.
One of these operating parameters is the air mass flow rate to the throttle valve. The air flow rate is sensed by a mechanical flap meter 11. The meter comprises a lightweight flap 12 pivotally mounted about an off-centre axis. A coil spring 13 is connected between a central bracket on the upstream surface of the flap 12 and the wall of. the manifold. The spring 13 biases the flap into the closed position shown in Figure 1 where it lies at 15° to a radialplane of the manifold. The flap opens to 90° to provide a 75° operating angle. The spindle of the flap is damped at one end by viscous fluid in a chamber. The other end of the spindle is connected to a rotary potentiometer having a voltage output which increases linearly with the angle of opening of the flap. Typically, the potentiometer produces a 5 volts output at an air flow rate of 12 lb/min.
A detailed description of the flap meter is included in the Applicants' co-pending EPC Patent Application No. 83307276.2.
The secondary primary parameters on whose value the fuel metering pulse is based is engine speed. The signal used to compute engine speed is obtained from the vehicle ignition circuit, for example, from the four vane switch of a Hall-effect distributor. After suitable pulse shaping a train of pulses 'ignition pulse signals' derived from the ignition pulses is fed to the control unit 20.
Other engine parameters are also sensed; these include the ambient air temperature and the engine cooling water temperature. The techniques for sensing these parameters are well known. A further relevant parameter is the voltage of the battery, which is sensed by a connection to the notional 12 volts rail. The actual battery voltage may vary between 8 and 18 volts depending upon battery condition and the numnber of connected electrical accessories.
A signal indicative of engine phase is also developed by an appropriate sensor and fed to the control unit 20. A suitable method of sensing engine phase is by attaching a vane to the shaft of the distributor to produce an electrical pulse once for each revolution of the engine camshaft. Typically this pulse coincide with the top-dead-centre crank position for cylinder No. 1.
A cranking-fail safe circuit based purely on hardware is connected to receive signals from, and transmit signals to, the electronic control unit. This circuitry is adapted to provide pulse width signals to the drive circuit of the injector valve during engine cranking or in the event of malfunction of the control unit 20.
In the cranking mode, this circuit produces large width pulses synchronised to the engine crnaking speed. This pulse width may be sequentially reduced over a period of say 20 seconds of cranking.
The fail-safe mode provides a limp home fuel supply whenever the microprocessor system fails. A single pulse width is produced with engine speed.
The microprocessor in the control unit activates the limp-home circuitry in the event of a detected failure, but constantly undergoes a self test routine and regains control from the limp-home circuit when the failure condition is removed.
Figure 1 shows schematically three data storage areas in a memory unit 21 included in the electronic control unit 20. These include a basic pulse width matrix 22 which receives the air flow signal from the flap meter 11 and the engine speed signal from the ignition circuit. The output from the basic pulse width matrix 22 is fed to a cyclic distribution matrix 23 receiving the engine phase signal which modifies the pulse width value from the matrix 21 to take into account manifold variations. The 'distributed pulse width' signal is then modified by the compensation factors derived from 'look-up' tables in the memory on the basis of the sensed values of air temperature, coolant temperature, battery voltage, etc.
Figure 2 illustrates in block schematic form the electronic control unit 20 and its input and output signals. The data processing capability of the unit is provided by an Intel 8051 microprocessor 24. Analog signals representing the sensed engine parameters of air flow rate, coolant temperature and battery voltage are received as inputs to an analog multiplexer 25. The single multiplexed output signal is then fed to an analog-to-digital converter 26 which converts the analog signal to an 8-bit digital signal. The digital signal is received by the input circuitry of the microprocessor 24 and this circuitry also receives, an another input, the ignition pulse signal indicative of engine speed.
The memory unit 21 in the control unit is a programmable read-only memory in which has been stored the matrices and look-up table of values for the injector pulse width signal.
One output port from the control unit is connected to the solenoid drive circuit controlling injector operation and through which connection the fuel pulse width signal is fed to the drive circuit.
Referring now to Figure 3, the air flow rate signal derived from the flap meter 11 typically has a variation with time as shown when engine loading is constant. The variation is approximatley sinusoidal with each peak corresponding to the suction stroke of one cylinder of the engine. The frequency of the air flow waveform is thus equal to the repetition rate of the ignition pulses from the distributor. The microprocessor derives a steady value for the air flow by executing a sample program on the input air flow signal at fixed intervals of 0.4 milliseconds and numerically integrating the results using a program based on Simpson's Rule or the Trapezoidal Rule. The period over which sampling is conducted and integration carried out is clocked by the ignition pulse signal. A sampling period is commenced at the beginning of each alternate pulse in the train of pulse constituting the ignition pulse signal and continues until the beginning of the following pulse. It will be appreciated from Figure 3 that such a sampling period covers a full cycle of the air flow waveform. It will be further appreciated that since sampling occurs at fixed intervals the number of samples obtained during one period is dependent upon engine speed.
The time between ignition pulses at maximum engine speed may be 5 milliseconds so that 12 samples of the air flow waveform are obtained from which to average and so derive a steady value, whereas at slowest speeds more than one hundred samples may be obtained between ignition pulses.
The value representing engine speed that is used to address the basic pulse width matrix 22 in the memory unit 21 is derived by a procedure carried out in association with the sampling of the air/flow waveform. The 8051 microprocessor 24 includes an 8-bit counter and this receives the 0..4 millisecond sampling pulses derived from the microprocessor clock running at 8 MHz. The counter is set oy tne . microprocessor on alternate ones of the ignition pulses and reset by the next ignition pulse. Thus the count output represents the number of 0.4 millisecond periods between a pair of ignition pulses and hence engine speed. It is desirable to obtain an accurate measurement of engine speed, preferably to a resolution of 6 revs per minute. It will be appreciated that at low speeds a large number of the fixed length sampling pulses are counted giving high resolution, which is particularly advantageous in the computation of fuel delivery when the engine is idling at at low speeds. The resolution is further improved using an additional 3-bit counter which is reset by each sampling pulse and counts timing pulses from a 20 KHz clock. It therefore divides each 0.4 millisecond sampling interval into 8 further sub-divisions, so that its count at the end of a sampling period represents the duration between the final sampling pulse and the end of the sampling period.
The accurate engine speed signal and the steady airflow signals are used to address the basic pulse width matrix 22 in the memory unit 21. This matrix comprises 256 values for the injector pulse width corresponding to 16 engine speeds and 16 airflow values. The injector pulse width values are programmed into the memory and are chosen on the basis of experimental determinations of the optimum values for given engine speeds and air flow rates. These values will vary for different types of vehicles and different engines. Typically the matrix axes comprises engine speed values in the range from 400 - 7000 RPM and _*ir flow rates in the range 0-12 lb/min. The stored values for the injector pulse widths are typically from 0.99 to 12 milliseconds. The increments between successive addressing values of the parameters need not be linear. In one example the addressing values for airflow rate are in a linear progression whereas those for engine speed are in a non-linear progression. Intermediate values for the injector pulse width are obtained by interpolation, the interpolation between matrix values may however be linear even when one or both of the axis of the matrix are non-linear.
The read-out value from the basic pulse width matrix 22 gives a basic pulse width value.
This pulse width is subjected to modifications before being fed to the injector. Firstly the pulse width signal is modified by a value obtained from the cyclic distribution matrix 23 also in the memory 21. This matrix is addressed by the accurate engine speed signal and the signal representing the phase of the engine cycle. The cyclic distribution matrix scales the pulse width signal from the basic fuel matrix by a factor in the range - 30% so that manifold variations to particular cylinders may be accommodated.
The resultant, modified pulse width, referred to as a 'distributed pulse width' is then further adjusted by a compensation factor derived from the sensed values of other engine operating parameters. Tables of compensation factors are stored in memory 21 for coolant temperature, air temperature, air absolute pressure and battery voltage. Each table typically has 8 addressing values for the parameter and an interpolation facility may be provided for values of the parameter between these addressing values. The coolant temperature table provides a multiplying factor for the pulse width value so that the fuel supply is enriched in cold start conditions. The tables for air temperature and air absolute pressure provide multiplying factors to compensate for ambient conditions. The battery voltage table provides an additive factor to compensate for variations in battery potential.
Provided that the engine is not subject to acceleration the compensated pulse width signal is fed to the solenc*-d drive circuit, to control the opening period of the ball value in the injector and hence the quantity of fuel delivered.
The sampling technique used to obtain a relatively noise-free air flow measurement and the accurate engine speed measurement is also used to detect acceleration. The opening of the throttle for acceleration by a driver produces a change in the air flow at the intake manifold more quickly than a change in the ignition pulse period. The microprocessor 24 compares successively sample-averaged values of the air flow, and if the ratio of the second value to the first exceeds a predetermined threshold, for example, 10%, a further compensating factor table in memory is enabled and the distributed pulse width is increased dependent on the ratio of the successive air flow values. This compensating enrichment factor modifies the pulse width to 100% - 300% of the distributed pulse width. The microprocessor is programmed to maintain the enrichment factor for a predetermined number of engine revolutions, typically 20 revolutions and then gradually to reduce the factor over a further predetermined number of engine revolutions, for example the subsequent 20 revolutions. In this manner an immediate large increase in fuel is supplied to the engines in response to the operation of the throttle. The optimum formula for an enrichment stragegy over a number of engine revolutions following detection of intended acceleration can be determined by experimentation and will vary from engine to engine.
Referring now to Figure 4, the sequence of operations carried out by the microprocessor 20 is shown in the form of a flow diagram. The program is set up by an initialisation of various pointers, putting the initial values of temperature, battery voltage, acceleration threshold etc. into the appropriate locations in memory. The program commences by checking status flaps to decide which of the two main sub-routines - Updating Parameters and Updating Output is to be carried out. In the Updating Parameters sub-routine the program checks whether the flag indicating the detection of acceleration has been set. If so, the enrichment strategy sub-routine is followed. Since the enrichment stragegy may require that the quantity of fuel be enriched over a number of engine cycles the acceleration flag may remain set for a number of cycles, and the enrichment stragegy sub-routine determines the amount of enrichment dependent upo! the amount of acceleration and the period for which the acceleration flag has been set.
Following the acceleration check, the updating of the main parameters takes place. The sampling of the mass flow waveform and the summation of the sampling pulses to determine engine speed take place in parallel, and the operation lasts for half of an engine revolution. i.e. between two ignition pulses. During this period the fuel pulse width signal is fed from the control unit to the injector drive circuit at the appropriate point in the cycle. The value of this signal is based on the existing values of the main parameters at the start of the update routine as modified by cyclic distribution matrix, the correction factors and the enrichment strategy.
Upon completion of the parameter updating the status flag is set such that upon return to the Check Status Flag stage the program next follows the Update Output sub-routine. Firstly an acceleration check is carried out as at the beginning of the Update Parameters Sub-routine. The accurate value of the engine speed is then updated from the 8-bit signal indicating the number of samples in the last period and the 3-bit indication of the fraction of the last sampling interval that preceded the ignition pulse. The basic fuel matrix in memory is then addressed and the read-out value is corrected by the cyclic distribution matrix and the compensation factors. The resultant fuel pulse value is then registered for use in the next injection pulse signal.
During the matrix interrogation steps a fuel pulse width signal is fed to the injector drive circuit at the appropriate point in the engine cycle.
Since the Update Output sub-routine takes less than half of one engine revolution to complete, a further acceleration check may be carried out at the end of this sub-routine. This may comprise the sampling of the mass flow signal for a number of intervals in the remaining period before the next ignition pulse. This is then compared with the previous value to detect acceleration.
The routine then returns to the Check Status Flag state where it is then redirected again to the Update Parameters sub-routine.
In another example, the flap meter used to detect air flow rate and hence a function of engine loading in the intake manifold is omitted. Instead a piezo-resistive pressure transducer is used to sense absolute manifold pressure which is another function of engine loading. The transducer may be located within the electronic control unit and communicate with the manifold by a length of tubing. The value of absolute manifold pressure follows a similar cyclic variation to that of the air flow past the flap meter, the two parameters being closely related. The values of absolute manifold pressure sensed by a transducer are however less subject to damping than the air flow values and the waveform is less uniformly sinusoidal. As a consequence it provides a more accurate and immediate indication of changes in engine loading. In an alternative embodiment of the invention, the signal representing engine loading, whether derived from air flow or absolute manifold pressure, is not used to detect acceleration. Instead a direct indication of throttle position is obtained from a potentiometer on the spindle of the accelerator pedal. This provides an immediate indication of the driver's depression of the pedal, and hence of a command to accelerate the vehicle. Such throttle potentiometers are well known and in a typical version it is provided with dc power; the position of the acceleration pedal modifying the position of a slider within the potentiometer to give a dc signal indicative of the degree of depression. This signal is fed to the electronic control unit via the analogue to digital converter, where the rate of change of the signal is computed by a sampling procedure.
In this system the values of engine loading and speed may be determined as in the above described embodiments, and these control the basic pulse width as before. The signal from the throttle potentiometer is sampled in the same manner as the pressure waveform, that is to say at 0.4 msec intervals between ignition pulses. Detection of acceleration may therefore be made after 0.4 msecs of the change in the signal from the potentiometer.
The enrichment strategy following detection of a change in throttle angle is immediately to set a multiplication factor, say 1.5, in a compensation table, and immediately to commence the output of a fuel metering pulse irrespective of synchronism with the ignition pulses. The multiplication factor is maintained for a number of successive fuel metering pulses which number is determined by the value of the engine loading (absolute manifold pressure or air flow rate). The multiplication factor is then reduced over a following number of fuel metering pulses, which number is again determined by the engine loading. The multiplication factor is also variable and is dependent on the determined rate of change of throttle angle. Thus a rapid depression of the acceleration pedal produces a greater multiplication factor than a slow depression.
The relationships between the rate of change of throttle and multiplication factor and between the number of metering pulses which are extended by the factor and the engine loading are determined from look-up tables, previously programmed into memory, from data obtained experimentally. Thus these relationships may vary from one type of vehicle to another.
Figure 5 illustrates graphically one case A where the rate of change of the throttle angle is relatively large and the multiplication factor is consequently also large, whereas in case B the rate of change is less. The load on the engine is however different in the two cases resulting in the lesser factor being applied over a larger number of metering pulses. It is to be noted that in both cases the enrichment linearly decreases to zero at some point during the procedure.
Figure 6 illustrates the sequence of operations carried out by the microprocessor in this enrichment procedure. The program is set up by initialisation of various pointers and status flags are set, as in the program described with reference to Figure 4. The program also similarly includes the two main sub-routines Up-dating Parameters and Updating Outputs, and most of these routines are common to the program described in Figure 4 and will not be described further. The significant difference is that following a check of the current acceleration status, if acceleration enrichment is not taking place an immediate check is made as to whether acceleration has been detected from the sampled throttle potentiometer signal, if it has then an immediate initial enrichment strategy takes place comprising the delivering of the asynchronous fuel metering pulse described above.
The control unit generally as described in the foregoing may, in another embodiment of the invention, be used to control an electronic carburetter. In this example fuel is injected into the carburetter float chamber via a ball valve injector controlled by an electrical pulse applied to its solenoid drive circuit. The injection pulses are timed in the control unit in relation to the ignition pulse signal which is an input to the control unit as in the previous example. It is important in this case that the injector valve opens during the suction stroke of each engine cylinder.
In a further example, the electronic control unit also controls the fuel supplied by an auxilliary liquid petroleum gas (LPG) system . Such an auxilliary system may be used in conjunction with a single-point fuel injection system as aforesaid, and each system may be selected by a driver's operating switch on the dashboard of the vehicle. The memory in the control unit has a further fuel supply matrix appropriate to the LPG system, and this may be addressed by the same primary parameters as in the single-point injection system. Selection of either fuel system by the driver causes the microprocessor to consult the appropriate memory matrix.

Claims

1. Electronic apparatus for controlling the supply of fuel to an internal combustion engine, the electronic apparatus being adapted to receive, in use, a first electrical signal representing a function of engine loading and a second electrical signal having, or representing, a periodicity derived from the cylinder firings of the engine, the electronic apparatus producing an output signal which, in use, is connected to control a fuel metering pulse, said output signal having a value selected or derived by said electronic apparatus on the basis of at least the value of a signal representing an averaged value of the function of engine loading, characterised in that the electronic apparatus includes a signal processing means operative to sample said first electrical signal at fixed intervals over a period fixed to the period of said second signal, and further operative to derive a signal representing an averaged value of said samples.

2. Electronic apparatus as claimed in claim 1 wherein said average value is derived from said samples by a numerical integration method determined by a computer program controlling said signal processing means.

3. Electronic apparatus as claimed in claim 2 wherein the numerical integration method is based on Simpson's Rule or a Trapezoidal Rule.

4. Electronic apparatus as claimed in any one of the preceding claims characterised by the inclusion of electronic counting means operative to count the number of sample values of said first electrical signal during one or more periods of said second electrical signal, and hence derive a signal representing engine speed, the value of said output signal from the electronic apparatus being selected or derived also on the basis of said signal representing engine speed.

5. Electronic apparatus as claimed in any one of the preceding claims wherein the signal processing means is further operative to compare sequential ones of said signals representing averaged values of the function of engine loading and to initiate an additional increase in the value of said output signal when said sequential signals are increasing in value above a predetermined amount of rate.

6. Electronic control apparatus for controlling the supply of fuel to an internal combustion engine including an input connected in use to receive a signal indicating intended or actual engine acceleration, the electronic control apparatus having an output at which, in use, a signal controlling a fuel metering pulse is produced, characterised in that the control apparatus is responsive to said signal indicating acceleration to increase the value of said output signal and hence the amount of fuel delivered to the engine, by a predetermined constant amount over a predetermined range of engine cycles and by a reduced amount over a subsequent range of engine cycles.

7. Electronic control apparatus as claimed in claim 6 characterised in that said predetermined constant amount by which the value of the output signal is increased is a proprtion of that value.

8. Electronic control apparatus as claimed in claim 6 or claim 7 wherein said predetermined constant amount is computed or selected on the basis of the degree of intended or actual engine acceleration.

9. Electronic control apparatus as claimed in any one of claims 6 to 9 wherein said predetermined range of engine cycles is computed or selected on the basis of an electrical signal representing engine loading.

10. Electronic control apparatus as claimed in any one of claims 6 to 9 in combination with means for detecting changes in throttle angle and producing an electrical signal representative thereof, said electrical signal being sampled at fixed intervals to provide said signal indicating intended engine acceleration.