JPH0719100A - Method for setting injection starting internal for each cylinder of multi-cylinder engine - Google Patents

Abstract

PURPOSE:To reduce a load of interruption process by dividing a single cycle of respective cylinders of a multi-cylinder engine into eight intervals, expressing these respective intervals by a bit 0 to a bit 7 as injection starting intervals, and setting the injection starting intervals of the respective cylinders in one or two more places according to the bit. CONSTITUTION:A single cycle of respective cylinders of a multicylinder engine is divided into eight intervals, and these respective intervals are expressed by a bit 0 to a bit 7 as injection starting intervals, and the bit corresponding to an injection starting interval setting flag IJTMGF is set in one or two more places, and the injection starting interval is specified. Whether or nor a present interval is a designated injection starting interval is deduced in interruption process carried out in synchronism with engine speed by making it correspond to the bit designated by the injection starting interval setting flag IJTMGF, and injection is started in the prescribed timing in the designated interval. Thereby, a load of the interruption process is reduced, and excellent controllability of fuel injection can be obtained.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for setting an injection start section for each cylinder of a multi-cylinder engine in which one cycle of each cylinder is divided into eight sections and an injection start section is designated.

[0002]

2. Description of the Related Art In recent years, microcomputers have been introduced into vehicles such as automobiles, and it has become possible to control engines, power trains and the like with high precision. As a result, the development of microcomputer software has become a major factor in the development of vehicle control systems.
Improving the efficiency of processing on control algorithms has become an important issue. Particularly, in recent engine control systems, the control content is complicated, and efficient utilization of the CPU is important for improving the engine controllability.

Conventionally, the fuel injection start timing in an engine control system using the above microcomputer is obtained by counting crank pulses input at constant crank angle intervals, as disclosed in, for example, Japanese Patent Application Laid-Open No. 61-160545. One that detects the rotational angle position of the engine and sets it based on the detected value (so-called angle control), or one that sets the injection start timing based on the interval time measured between predetermined crank angles (so-called time control), etc. is there.

[0004]

By the way, when performing sequential control in a multi-cylinder engine, conventionally,
In accordance with the instruction to start injection from before the start of the intake stroke, the current crank angle is calculated several times before the intake stroke in the interrupt process synchronized with the engine rotation, and the result is compared with the instruction value for injection. The determination as to whether to start is repeated for all cylinders.

Further, in the engine for controlling the injection end, the injection end timing is calculated from the current engine speed,
The procedure for back-calculating the injection start timing from this end timing is performed during the above-described interrupt processing for each crank angle.

As a result, in the fuel injection control, the interrupt processing in synchronism with the engine rotation is complicated and heavy, and the injection control cannot be performed sufficiently. This problem is particularly remarkable in a control system that executes multiple injections in one cycle.

The present invention has been made in view of the above circumstances, and reduces the load of interrupt processing executed in synchronization with engine rotation, and multi-cylinder in which the load of interrupt processing is small even when injection control is performed a plurality of times. It is an object of the present invention to provide a method of setting an injection start section for each cylinder of an engine.

[0008]

In order to achieve the above object, a method for setting an injection start section for each cylinder of a multi-cylinder engine according to the present invention is such that one cycle of each cylinder of the multi-cylinder engine is used in the latter half of the intake stroke, the first half of the intake stroke, and the exhaust gas. Second half of the stroke, first half of the exhaust stroke,
It is divided into 8 sections of the latter half of the combustion stroke, the first half of the combustion stroke, the latter half of the compression stroke, and the first half of the compression stroke. Each of these sections is represented by bit 0 to bit 7 as an injection start section, and is designated by one or more designated bits. The injection start section of each cylinder is set based on this.

[0009]

[Operation] In the present invention, one cycle is divided into eight sections,
Each section is assigned to bit 0 to bit 7, and in the interrupt processing synchronized with the engine rotation, it is determined for each cylinder whether the current section is the designated injection start section based on the designated bit at one or two or more locations. to decide.

As a result, the injection start section can be specified only by setting or clearing an arbitrary bit portion of bit 0 to bit 7, and even if the injection is executed a plurality of times in one cycle, the interrupt processing is performed. This burden is reduced.

[0011]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 relates to an embodiment of the present invention, and FIG. 1 is a flow chart showing a periodic interrupt process every 0.5 ms.
FIG. 2 is a flowchart showing Krasen interrupt processing, and FIG.
Is a flowchart showing a job priority process, FIGS. 4 to 7 are flowcharts showing a job execution subroutine, FIG. 8 is a flowchart showing a crank position calculation subroutine,
9 is a flowchart showing a CCAS / RCAS discrimination subroutine, FIGS. 10 and 11 are flowcharts showing a fuel injection start timing calculation routine, FIG. 12 is a flowchart showing an injection timer setting subroutine, FIG. 13 and FIG.
4 is a flow chart showing the injection timer set macro 1 routine, FIG. 15 and FIG. 16 are flow charts showing the injection timer set macro 2 routine, FIG. 17 is an explanatory diagram showing a job execution state, FIG. 18 is an explanatory diagram of a job flag, FIG. 1
9 is an explanatory diagram showing changes in the job execution flag and the overlap counter, FIG. 20 is an explanatory diagram of the system shift buffer, FIG. 21 is an explanatory diagram of the class interval table, and FIG.
24 is an explanatory view of a cylinder / crank position state map, FIG. 23 is a crank position, a cam position and a crank position variable, an injection start section table, and a time chart in the order of strokes by cylinder, FIG.
Is an explanatory view of an injection start section and an injection start section setting flag, FIG. 25 is a time chart showing injection timer control, FIG. 26 is a schematic configuration diagram of an engine system, FIG. 27 is a front view of a crank rotor and a crank angle sensor, and FIG. 28 is a front view of the cam rotor and the cam angle sensor, and FIG. 29 is a circuit configuration diagram of the electronic control system.

In the engine control system of this embodiment, the engine system shown in FIG. 26 is controlled by an electronic control unit (ECU) 50 having a microcomputer shown in FIG. 29 as a core, and fuel injection control, ignition timing control and the like are performed. Be seen. The microcomputer of the ECU 50 is equipped with an operating system (OS) based on a new concept, and under the control of this OS, each job based on each control strategy is executed according to each priority level.

First, the equipment configuration of the engine system will be described.

In FIG. 26, reference numeral 1 is an engine (horizontally opposed four-cylinder engine in the present embodiment), # 1 and # 3 cylinders are on the right bank R side, and # 2 and # are on the left bank L side.
Equipped with 4 cylinders. Then, the intake ports 2a of the cylinder heads 2 provided on both the left and right banks of the cylinder block 1a.
An intake manifold 3 is communicated with the intake manifold 3, and a throttle passage 5 is communicated upstream of the intake manifold 3 via an air chamber 4. An air cleaner 7 is attached to the upstream side of the throttle passage 5 via an intake pipe 6, and the air cleaner 7 is in communication with an air intake chamber 8 which is an intake port for intake air.

An exhaust pipe 10 is connected to the exhaust port 2b via an exhaust manifold 9, and a catalytic converter 11 is connected to the exhaust pipe 10 and is connected to a muffler 12. On the other hand, a throttle valve 5a is provided in the throttle passage 5, and an intercooler 13 is provided in the intake pipe 6 immediately upstream of the throttle passage 5,
Further, a resonator chamber 14 is interposed downstream of the air cleaner 7 in the intake pipe 6.

Further, an idle speed control valve (IS) is provided in a bypass passage 15 which connects the resonator chamber 14 and the intake manifold 3 and bypasses the upstream side and the downstream side of the throttle valve 5a.
CV) 16 is interposed. Furthermore, this ISCV1
A check valve 17 which is opened immediately downstream of 6 when the intake pressure is a negative pressure and which is closed when the intake pressure becomes a positive pressure by being supercharged by a turbocharger 18 is interposed.

In the turbocharger 18, a compressor is installed downstream of the resonator chamber 14 of the intake pipe 6, and a turbine is installed in the exhaust pipe 10. Further, a wastegate valve 19 is provided at the turbine housing inlet of the turbocharger 18, and a wastegate valve actuating actuator 20 is connected to the wastegate valve 19.

The waste gate valve actuating actuator 20 is partitioned into two chambers by a diaphragm, one of which forms a pressure chamber communicating with the waste gate valve controlling duty solenoid valve 21, and the other of which forms the waste gate valve 19. A spring chamber is formed that houses a spring that urges in the closing direction.

The wastegate valve controlling duty solenoid valve 21 is provided in a passage that connects the resonator chamber 14 and the compressor downstream of the turbocharger 18 of the intake pipe 6, and is controlled by the ECU 50. The pressure on the resonator chamber 14 side and the pressure on the compressor downstream side are adjusted according to the duty ratio of the signal, and the pressure is supplied to the pressure chamber of the waste gate valve operating actuator 20.

That is, the wastegate valve controlling duty solenoid valve 21 is controlled by the ECU 50, and the wastegate valve operating actuator 20 is controlled.
Is operated to adjust the exhaust gas relief by the waste gate valve 19, so that the supercharging pressure by the turbocharger 18 is controlled.

An absolute pressure sensor 22 is connected to the intake manifold 3 through a passage 23, and the passage 2
3, an intake pipe pressure / atmospheric pressure switching solenoid valve 24 that selectively communicates the absolute pressure sensor 22 with the intake manifold 3 or the atmosphere is interposed.

Further, an injector 25 is provided immediately upstream of each intake port 2a of each cylinder of the intake manifold 3.
In addition, an ignition plug 26a whose tip is exposed to the combustion chamber is attached to each cylinder of the cylinder head 2, and an igniter 27 is connected to an ignition coil 26b connected to the ignition plug 26a. There is.

The injector 25 includes a fuel tank 2
Fuel is pressure-fed from an in-tank type fuel pump 29 provided inside 8 through a fuel filter 30, and the pressure is regulated by a pressure regulator 31.

An intake air amount sensor 32 is provided in the intake pipe 6 immediately downstream of the air cleaner 7, and a throttle sensor 33 having a throttle opening sensor 33a and an idle switch 33b built in the throttle valve 5a. Are lined up.

Further, a knock sensor 34 is attached to the cylinder block 1a of the engine 1, and a cooling water temperature sensor 36 is exposed to a cooling water passage 35 which connects the left and right banks of the cylinder block 1a to the exhaust pipe 10. O2 at the collecting part of the above exhaust manifold 9
The sensor 37 is exposed.

A crank rotor 38 is rotatably mounted on the crank shaft 1b supported by the cylinder block 1a, and a crank angle sensor 39 including an electromagnetic pickup is provided on the outer periphery of the crank rotor 38. Further, a cam angle sensor 41 for discriminating a cylinder, which is composed of an electromagnetic pickup or the like, is provided opposite to a cam rotor 40 connected to the cam shaft 1c of the engine 1. still,
The crank angle sensor 39 and the cam angle sensor 41
Is not limited to a magnetic sensor such as an electromagnetic pickup, but may be an optical sensor.

As shown in FIG. 27, the crank rotor 38 has protrusions 38a, 38b and 38c formed on the outer periphery thereof, and these protrusions 38a, 38b and 38c are associated with respective cylinders (# 1, # 2 and # 2). Before # 3, # 4 compression top dead center (BTD
C) It is formed at the positions of θ1, θ2, and θ3. In this embodiment, θ1 = 97 ° CA, θ2 = 65 ° CA,
θ3 = 10 ° CA.

Each protrusion of the crank rotor 38 is detected by the crank angle sensor 39, and as shown in FIG. 23, a crank angle signal (class signal) of BTDC 97 °, 65 °, 10 ° CA is 1/2 revolution of the engine. Every (18
Output every 0 ° CA). Then, the input interval time of each signal is counted by a timer, and the engine speed is calculated. The protrusion 38b serves as a reference crank angle at the time of setting the ignition timing, and the protrusion 38c serves as a reference crank angle at the start-time injection start timing and a crank angle indicating a fixed ignition timing at the start.

Further, as shown in FIG. 28, on the outer periphery of the cam rotor 40, there are projections 40a, 40b for cylinder discrimination,
40c is formed, and the protrusion 40a is formed at the position after compression top dead center (ATDC) θ4 of the # 3 and # 4 cylinders.
b is composed of three protrusions, and the first protrusion is A for cylinder # 1.
It is formed at the position of TDCθ5. Furthermore, the protrusion 40
c is formed by two protrusions, and the first protrusion is A for cylinder # 2.
It is formed at the position of TCDθ6. In this embodiment, θ4 = 20 ° CA, θ5 = 5 ° CA, θ6 = 2
It is 0 ° CA.

When each projection of the cam rotor 40 is detected by the cam angle sensor 41 and the combustion stroke sequence of each cylinder is # 1 → # 3 → # 2 → # 4, this combustion stroke sequence and Cylinder discrimination is performed based on a pattern (see the time chart of FIG. 23) of the detection signal of the cam angle sensor 41 and the value counted by the counter.

On the other hand, as shown in FIG. 29, the ECU 50
Is mainly composed of two computers, a main computer 51 for performing fuel injection control, ignition timing control, and the like, and a dedicated sub computer 52 for performing knock detection processing. A constant voltage that supplies a predetermined stabilizing power source to each part. The circuit 53 and various peripheral circuits are incorporated.

The constant voltage circuit 53 includes an ECU relay 54.
The relay coil of the ECU relay 54 is connected to the battery 55 via an ignition switch 56.
Further, the constant voltage circuit 53 is directly connected to the battery 55, and further, the fuel pump 29 is connected via a relay contact of a fuel pump relay 57.

That is, the constant voltage circuit 53 supplies control power when the ignition switch 56 is turned on and the relay contact of the ECU relay 54 is closed, and the ignition switch 56 is turned off.
Supply the power for backup when

The main computer 51 has a CPU 5
8 (hereinafter, referred to as main CPU 58), ROM 59,
RAM 60, the ignition switch 56 is OFF
Backup RAM 6 which holds data by being supplied with backup power from the constant voltage circuit 53 even when
1, a counter / timer group 62, a serial communication interface SCI 63, and an I / O interface 64 are connected to each other via a bus line 65.

The counter / timer group 62 includes various counters such as a free-running counter, a cam angle counter (hereinafter, abbreviated as "CamSen") signal counting counter, an injection timer, an ignition timer, and the like. A periodic interrupt timer for generating a periodic interrupt every 0.5 ms, a crank angle sensor (hereinafter, abbreviated as "classen") signal input interval timing classen timer, and a watchdog timer for system abnormality monitoring, etc. The various timers are collectively referred to for convenience, and other various software counters and timers are used in the main computer 51.

In the sub computer 52, the CPU 71 (hereinafter referred to as the sub CPU 71), the ROM 72, the RAM 73, the counter / timer group 74, the SCI 75, and the I / O interface 76 are also connected to the bus line as in the main computer 51. The microcomputer is a microcomputer connected via 77, and the main computer 51 and the sub computer 52 are the SCI 63, 7
They are connected to each other via a serial communication line via 5.

The I / O interface 64 of the main computer 51 has an intake port, an intake air amount sensor 32, a throttle opening sensor 33a, and a water temperature sensor 3 at its input ports.
6, O2 sensor 37, absolute pressure sensor 22, vehicle speed sensor 4
2 and battery 55 are A / D of 8 channel input
The idle switch 33b, the crank angle sensor 39, and the cam angle sensor 41 are connected together via the converter 66.
Is connected, and further, a starter switch 43 is connected to detect the starting state.

In this embodiment, the A / D converter 66 uses inputs for 7 channels, and the remaining 1
The channel is reserved.

An igniter 27 is connected to the output port of the I / O interface 64, and further,
Through the drive circuit 67, the ISCV16, the injector 2
5, the relay coil of the fuel pump relay 57, the waste solenoid valve controlling duty solenoid valve 21, and the intake pipe pressure / atmospheric pressure switching solenoid valve 24 are connected.

On the other hand, the I / O of the sub computer 52
The interface 76 has an input port to which the crank angle sensor 39 and the cam angle sensor 41 are connected, and
The knock sensor 34 is connected via the D / D converter 78, the frequency filter 79, and the amplifier 80. The knock detection signal from the knock sensor 34 is amplified to a predetermined level by the amplifier 80, and then the frequency filter 79. The necessary frequency components are extracted by the A / D converter 78.
At, it is converted into a digital signal and input.

The main computer 51 processes the detection signals from the sensors and calculates the fuel injection pulse width, the ignition timing and the like. That is, the intake air amount sensor 32
The intake air amount is calculated from the output signal, and the fuel injection amount corresponding to the intake air amount is calculated based on various data stored in the RAM 60 and the backup RAM 61, and the ignition timing and the like are calculated.

Then, a drive pulse width signal corresponding to the fuel injection amount is output to the injector 25 of the corresponding cylinder at a predetermined timing via the drive circuit 67 to inject fuel, and at the predetermined timing, the igniter 27 is also used. To the ignition plug 26a of the corresponding cylinder.

As a result, the air-fuel mixture supplied to the corresponding cylinder explodes and burns, the oxygen concentration in the exhaust gas is detected by the O 2 sensor 37 facing the collecting portion of the exhaust manifold 9, and the waveform of this detection signal is shaped. After that, the main CPU 58 compares it with a reference value (slice level) to determine whether the air-fuel ratio state of the engine is on the rich side or the lean side with respect to the target air-fuel ratio, and the air-fuel ratio is set to the target air-fuel ratio. Feedback control is performed so that

On the other hand, in the sub computer 52, a sample section of the signal from the knock sensor 34 is set based on the engine speed and the engine load, and the signal from the knock sensor 34 is A / D at high speed in this sample section. Then, the vibration waveform is faithfully converted into digital data to determine whether knock has occurred.

The output port of the I / O interface 76 of the sub computer 52 is connected to the input port of the I / O interface 64 of the main computer 51, and the knock determination result of the sub computer 52 is I / O. It is output to the interface 76. Then, in the main computer 51, when the sub computer 52 outputs a determination result indicating that knock has occurred, the knock data is read from the serial communication line via the SCI 63, and the ignition timing of the corresponding cylinder is immediately read based on the knock data. Delay and avoid knocks.

In such engine control, in the main computer 51, various programs for each item such as signal input processing from each sensor, engine speed calculation, intake air amount calculation, fuel injection amount calculation, ignition timing calculation, etc. Is efficiently executed under the control of one OS. This OS has various management functions for vehicle control and an internal strategy closely related to this management function, and systematically combines various jobs.

As the management function of the OS, (1-1) priority processing of jobs (1-2) divided file support for each job by section definition (1-3) stack usage status monitoring function (1-4) Abnormal interrupt operation monitoring function (1-5) Functions such as standard map and standard work memory settings that do not create unique restrictions for each job improve the control strategy development environment and maximize the limited CPU power. The same time interval processing which is the basis of digital control theory can be achieved as much as possible.

As the equal time interval processing, 2, 4, 10, 50, 250 are based on a periodic interrupt every 0.5 ms.
Five types of equally interrupted jobs for every ms are prepared, and a high-priority class job (hereinafter simply referred to as classen job) that is immediately interrupted by crank angle signal input as processing synchronized with engine rotation. , A relatively low-urgency low-priority job that is interrupted by a crank angle signal input when there is no other job with a higher priority.

Each of these jobs includes a Krasen job>
A priority level of 7 to 1 is assigned from a high side to a low side in the order of 2 ms job> 4 ms job> 10 ms job> low priority class job> 50 ms job> 250 ms job, and as shown in FIG. The low-speed job is divided into the high-speed job and processed, and the multiplex waiting process of each job is performed.

Each program operating under the OS is arranged in order for each function-specific area, and each function is named by a section declaration. The various strategy files (user files) that work under the OS use section declarations with the same name as the OS for each function, so that at the development stage, for example, initial value setting processing, 10 ms processing, background Even if the processes and the like are described in separate files, when the files are linked, the same processes are collected in one continuous area and integrated with the OS to operate integrally.

The main section areas used on each strategy file side are: ○ variable declaration area ○ self file name, automatic recording area at file creation ○ setting data area ○ class job area ○ 2ms job area ○ 4ms job area ○ 10ms Job area ○ Low priority class job area ○ 50ms job area ○ 250ms job area ○ Reset initialization job area ○ Initialization job area at stall ○ Background job area ○ This is the area of the program body and divides the file for each function It enables program development and structured description of programs.

Further, the above-mentioned OS has (2-1) A / D conversion processing (2-2) Calculation of various information related to crank position (2-3) Debug as an internal strategy closely related to the above management function. Simulation function (engine rotation and A / D conversion) (2-4) Ignition timer set (2-5) Injection timer set and other functions are provided, and various service routines related to these functions are provided. It is prepared in the job.

Conventionally, such a function has been achieved at each job level, but in this system,
If the fuel injection amount, ignition timing, etc. are set for each job on the user side based on the A / D conversion result processed by the OS side, crank position information, engine speed, etc. The instruction value is set by the OS in the injection timer and the ignition timer.

Next, the function of fuel injection control in the main computer 51 will be described with reference to the flowcharts shown in FIGS. Since the sub computer 52 is a computer dedicated to knock detection processing, its operation description is omitted.

First, the ignition switch 56 is turned on.
When the system is powered on, a reset interrupt associated with the reset is activated, various initializations are performed, and a periodic interrupt timer for activating the periodic interrupt is activated every 0.5 ms.
Each signal input from 9 (BTDC 97 °, 65 °, 10 °
The classen interrupt that is activated every six engine revolutions per CA) is permitted, and then the background job is executed.

Then, on this background job, a periodic interrupt every 0.5 ms and 6 times per engine revolution.
A job of the 7th level is preferentially processed by the number of classen interrupts. In these two interrupts, after executing their own processing, they jump to a common address and execute job priority processing.

The reset interrupt may be caused by a factor that does not normally occur, such as when division by 0 is executed in an internal operation or when an infinite loop occurs.
Is activated.

First, the periodic interrupt every 0.5 ms shown in FIG. 1 will be described. In this periodic interrupt, step S1
At 00, set the OS work area, and at step S101,
When the watchdog timer is initialized, the process proceeds to step S102, and the P-RUN flag is set once every 20 times, that is, 10 times.
Invert every ms. The P-RUN flag is a flag for preventing the system from being automatically reset by a protection circuit (not shown), and is protected as long as the system operates normally and is inverted at regular time intervals (every 10 ms). The operation of the circuit is blocked.

Next, in step S103, the switch output is transferred. This switch output is the ON / OFF value of the bit written in the memory in each job. It is not directly output from the output port of the I / O interface 64 from each job, but the OS outputs the memory every 0.5 ms. Transfer the value to the output port.

Next, in step S104, an A / D conversion subroutine is executed to make various settings related to A / D conversion, and in step S105 a job flag creation subroutine is executed to set 2, 4, 10 After creating the job flag JB_FLG indicating each job interruption request every 50, 250 ms, the A / D conversion is started in step S106.

In the A / D conversion, basically, the 8-channel input of the A / D converter 66 is processed every 0.5 ms in a predetermined conversion order, and all the inputs are converted in a cycle of 4 ms. However, one specific channel is synchronized at every crank angle of 90 ° (with time accuracy of 0.5 ms) for A / D conversion of suction pipe pressure that causes rotational pulsation, and interrupts the conversion order. The processing is performed in a dull shape, and the order of subsequent inputs is delayed by one.

When the engine speed is 3750 rpm or more, the last input of A / D conversion is completely stopped, and
At 500 rpm or more, the penultimate input is also stopped, but the order of A / D conversion is that the throttle opening, intake air amount, etc. that have a fast change, and the cooling water temperature, voltage, etc., that have a relatively slow change. Is set so that
Moreover, since the last A / D conversion order is set to the crank synchronization input, no particular trouble occurs.

Further, as shown in FIG. 18, the job flag JB_FLG is one in which each bit of the 1-byte variable is assigned as a flag corresponding to each job, and a plurality of job requests can be made simultaneously. There is. Bits 1 to 7 of this 1-byte variable correspond to priority levels 1 to 7, and are respectively 250 ms job, 50 ms job,
It is assigned to flags of low priority class job, 10 ms job, 4 ms job, 2 ms job, and class job. Then, when a predetermined bit is set, a job interruption request of a corresponding priority level is made. still,
Bit 0 is assigned to the flag of the background job and is not normally referred to.

Then, in step S105, the job flag JB_FLG is created by the job flag creation subroutine, and after A / D conversion is started in step S106, the process proceeds to step S107, and any job of the job flag JB_FLG is processed. Check whether the bit to be set is set.

As a result, when none of the bits of the job flag JB_FLG is set, there is no request from any job and the interrupt is ended.
If any bit of LG is set, the process proceeds to step S108, and there is no flag below the current level (a state in which a job of a predetermined priority level is being executed at the time when this periodic interrupt is executed). To find out.

If there is no flag below the current level in step S108, the process jumps to the job priority process shown in label WAR_JB in FIG. 3, and if there is a flag below the current level, in step S109 the level below the current level is reached. The overlap counter OLC of is increased by one.

The overlap counter OLC is a counter for storing a job request, 1 byte is allocated to each priority level, and the job flag JB_F is set.
It is incremented when a job is requested by LG and decremented when the job is completed. That is, by storing the job request by the counter, it is possible to deal with the multiple job request.

Then, the above steps S109 to S1
Go to step 10 to check whether there is a flag higher than the current level. If there is no flag higher than the current level, exit the routine and end the interrupt. If there is a flag higher than the current level, the job priority process of label WAR_JB is executed. Jump to.

On the other hand, in contrast to the periodic interrupt every 0.5 ms, the interrupt by the classen in FIG.
Then, when the OS work area is set, step S201
Then, the crank position calculation subroutine to be described later is executed, and the crank position variable for determining the current crank position and the elapsed time of the engine half rotation, that is, the half rotation time, which is the sum of the latest three Classen intervals, are calculated. To do.

The above crank position variables are system variables prepared in the OS, and as shown in FIG.
The crank position for the # 4 cylinder is determined by the crank angle sensor 3
BTDC 97 °, 65 °, 10 by signal input from 9
It is divided into 12 states by ° CA and represents the current crank position.

That is, for each cylinder, the crank position information variable S_CC indicating the Krasen input order by the numerical value of 0, 1, 2
AS, # 1 cylinder is 0, # 3 cylinder is 1, # 2 cylinder is 2, #
Cylinder information variable S_ indicating the combustion order of cylinders with 4 cylinders as 3
RCAS and 3 of the total position variable S_ACAS that comprehensively represent the Krasen order and the cylinder order with numerical values of 0 to 11
The current crank position is represented by a variable, and when the crank position is normally determined with certainty, it is set to 0,
The estimated state where the determination result does not match and the anxiety remains 1,
An error level S_ECAS that sets the unknown state to 2 indicates the determination situation of the crank position. 23, S_ indicating a system variable is omitted.

Next, from step S201 to step S2
When the process proceeds to 02, it is checked whether or not the crank position determination is normally completed or cannot be determined in the crank position calculation subroutine by reading the code stored in the accumulator A (error code 1, normal end code 0 ).

Then, in step S202, when the value of the accumulator A is 1 and the crank position cannot be determined, the interruption is ended, the value of the accumulator A is 0, and the crank position is determined normally. If so, the process proceeds to step S203, and the engine stall flag is cleared.

The engine stall flag is a flag indicating that the engine is in an engine stall state. When the classen interval is 0.5 sec or more (about 30 rpm or less), an engine stall processing routine prepared for a 50 ms job is provided. Is set by, and is cleared by this classen interrupt to release the stalled state.

Next, in step S204, the ignition timer setting subroutine is executed to set the ignition timer according to the ignition schedule created based on the instruction value of the ignition timing set by the user job. This ignition schedule is a structure variable having members such as dwell start timing, dwell on waiting time, and dwell off waiting time, and a preparation routine is prepared during a 10 ms job, and an ignition sequence is determined according to this ignition schedule.

Next, at step S205, an injection timer set subroutine described later is executed to set the injection time corresponding to the required fuel injection width for each cylinder set on the user job side.
It is calculated as an effective injection width per injection corresponding to single injection or multiple injections, is set in the injection timer, and the routine proceeds to step S206.

In step S206, it is determined whether or not the current job level at which this class is executed is the own job level. If the current status is the level of the class job, the class job is executed in steps S207 and S208. ,
Overlap counter OLC for low priority class job
Is incremented by 1 to terminate the interruption, and when the current job level is not the level of the class job, it is checked in step S209 whether the current job level is the level of the low priority class job or more.

If the current job level is equal to or higher than the low-priority class job, the process proceeds from step S209 to step S210, and if the overlap counter OLC of the low-priority class job is incremented by 1, step S2
At 11, the job flag of the class job is set, and the process jumps to the job priority process of the label WAR_JB.

On the other hand, if it is determined in step S209 that the current job level is not higher than the low priority class job, the process proceeds from step S209 to step S212 to set the job flag of the class job. The job flag is set, and the process jumps to the job priority process with the label WAR_JB.

In the job priority processing shown in FIG.
If the job level JB_LEV, which is a 1-byte variable indicating the priority level of the job, is incremented by 1 in S300, step S3
Go to 01 and check if the job flag corresponding to this priority level is set. If the job flag is not set, the process returns to step S300 to further increase the job level JB_LEV by one, and if the job flag is set, the process proceeds to step S302 to set the overlap counter OLC of the job for which the job flag is set. The initial value is changed from 0 to 1, and the process proceeds to step S303.

In step S303, it is checked whether or not there is a higher job flag. When there is a higher job,
Returning to step S300, the above processing is repeated. If there is no higher job, the process proceeds to step S304 to set the job executing flag JB_RUN, and then step S3
At 05, the highest-level job is executed by the job execution subroutine described later.

The job executing flag JB_RUN is
It is a flag that is set at the start of job execution and cleared at the end. By this flag, a job interrupted by a job with a higher priority can be identified during processing.

For example, as shown in FIG. 19, JB_LE
When an interrupt request for a 2ms job with JB_LEV = 6 is made during execution of a 10ms job with V = 4, the processing of the 10ms job is interrupted, and the higher priority 2ms job is set to JB_RUN = 1 and OLC = 1. Is executed. Then, during the processing of this 2 ms job, JB_
When an interrupt request for a 4 ms job with LEV = 5 occurs,
This 4 ms job is set to JB_RUN = 0 and OLC = 1 to accept an interrupt, but is not executed and is in a standby state.

After that, when the execution of the job by the job execution subroutine is completed, the process proceeds from step S305 to step S306, the overlap counter OLC is decremented by 1, and it is checked in step S307 whether the overlap counter OLC has become zero. . As a result, when the overlap counter OLC is not zero and there are multiple job interrupt requests at the same priority level, step S3
Returning to 05, the job is repeatedly executed, and when the overlap counter OLC becomes zero, the process proceeds from step S307 to step S308 to clear the job executing flag JB_RUN.

Next, in step S309, when the job level JB_LEV is decreased by 1 and the job level is moved to the next job level,
In step S310, it is checked whether this job level JB_LEV has become zero. Then, the job level JB_
When LEV is zero, this interrupt is ended, and when job level JB_LEV is not zero, step S3
Proceed to 11 and check whether the overlap counter OLC is zero.

When the overlap counter OLC is zero in step S311, there is no job request at this level. Therefore, the process returns from step S311 to step S309, the job level JB_LEV is further lowered by one, and the same processing is repeated. If the overlap counter OLC is not zero, the flow advances to step S312 to check if the job executing flag JB_RUN is set at this job level.

When the job executing flag JB_RUN is set in step S312, the job was being executed before the interruption, so the interruption is terminated and the process returns to the job before the interruption, and the job executing flag JB_RUN is set. If not, the process returns to step S304, the job of this level is executed, and the same processing is repeated.

That is, in FIG. 19, JB_LEV
= 6 ms 2ms job is completed, OLC = 0, JB_RU
When N = 0, the job level is lowered by 1 and JB_
4ms job with LEV = 5, JB_RUN = 0, OL
JB_RUN = 1 is set from the standby state of C = 1,
To be executed. Furthermore, when the 4ms job is completed, JB
Moving to _LEV = 4, the processing of the 10 ms job suspended by the 2 ms job and the 4 ms job is resumed from the state of JB_RUN = 1 (job is being executed).

In this way, the periodic interrupt every 0.5 ms,
A job flag J for notifying the priority level and execution timing of each job with the classen interrupt as the basic timing.
Since B_FLG is created, equal time interval processing and engine rotation synchronization processing can be realized as accurately as possible, and each job can be processed efficiently. Further, the job flag JB_F updated for each interrupt which is the basic timing
Since the overlap counter OLC stores the job multiplex request regardless of LG, even if the processing time of a certain job is prolonged and it is time to execute the same job again, the processing is abandoned midway. No, the process can be continued to the end as much as possible.

Next, the job execution subroutine of FIGS. 4 to 7 will be described.

First, in step S500, the job flag JB
By referring to _FLG, it is checked whether the job to be executed is not a class job. If it is not a class job,
The processing branches to label ALJ10, and if it is a class job, the process proceeds to step S501 to check whether or not cylinder discrimination is provided.

When the cylinder is not discriminated, the routine is not executed and the job is not executed. When the cylinder is discriminated, the steps from step S501 to step S501 are executed.
In step S502, the value of the overlap counter OLC is referenced to check whether or not the multiplex wait state is set.

In step S502, when the multiplex waiting state is not set, the process proceeds to step S503, and the system variable S_ACAS (crank total position variable) calculated for each classen interrupt is set as the user variable ACAS. On the other hand, in the multiplex waiting state, It branches to step S504, and user variable ACA
S is incremented by 1 and divided by 12, and the remainder is taken as a new user variable ACAS. This user variable ACAS is set to 0,
1, 2, ... 11, 0, 1, ... are updated by software.

That is, the classen job and the low-priority classen job may be delayed by being disturbed by the job itself or the job of high priority, but the classen interrupt is executed exactly in synchronization with the crank angle sensor signal, and the system variable S
_ACAS is updated regardless of job delay.

Therefore, in the job, the system variable S_AC
Even if the user knows the information about the cylinder and the crank position by referring to AS and tries to perform the work according to this information, if he / she is delayed by another job, the cylinder corresponding to his / her work And, it becomes impossible to know the information related to the crank position. Therefore, in the class job and the low-priority class job, the system variable S_ACAS for the OS is taken in as the user variable ACAS when not in the multiplex waiting state, and this user variable ACAS is updated for each job execution, and in the case of multiple request. Also, the information regarding the cylinder and crank position corresponding to itself is obtained so that appropriate processing can be performed.

After that, the process proceeds from step S503 or step S504 to step S505 to set the work area of the job. In step S506, the level 0 interrupt is permitted. In step S507, the classen job section is entered. Then, the processing linked to this class job section is executed, and in step S508, the interrupt is prohibited and the routine exits.

Next, at step S500, if the job to be executed is not a class job, the label AL
In step S510 of J10, it is checked whether or not the job is a 2 ms job. When the job is a 2 ms job, the work area of the job is set in step S511. In step S512, the level 0 interrupt is permitted, and in step S513, the 2 ms job Go to section. Then, the main body of the job (a routine based on the control strategy on the user side or a service routine prepared on the OS side) linked to this section is executed, and in step S514, interruption is prohibited and the routine exits.

On the other hand, if the job to be executed is not a 2 ms job in step S510, the process branches from step S510 to step S520 to check whether the job to be executed is a 4 ms job. If it is not a 4 ms job, the process branches to label ALJ30. If it is a 4 ms job, the work area of the job is set in step S521, and the process proceeds to step S522. Note that this 4ms job is
A / D conversion use job, which uses A / D conversion data via a system shift buffer SSHB described later.

In step S522, the interrupt of level zero is permitted, and then in step S523, the switch input is read, and in step S524, the section of the 4 ms job is moved to and the linked job main body is executed. After that, when the section of the 4 ms job is exited, the interrupt is prohibited in step S525, the process proceeds to step S526, the system shift buffer SSHB is shifted, and the routine exits.

The system shift buffer SSHB is
As shown in FIG. 20, the head offset addresses 0, +8, +1 in which the respective A / D conversion results of 8 channels are stored
It consists of memories of Nos. 6, +24, +32, +34, +36, and +38, and a 1-word memory at the start offset address -2 where the A / D conversion result of crank synchronization at every 4 ms is stored. One A / D conversion result executed each time is stored in one word (2 bytes).

From the start offset address 0, 4
It is a multi-stage shift memory, and A / D every 90 ° CA.
The conversion result is stored and the latest 4 data (1 rotation) can be referred to from the job. The head offset addresses +32, +34, +36, and +38 are each 1
This is a word memory, and when the smoothing function is selected, a weighted average value of the A / D conversion results is stored to enable noise removal and accuracy improvement. The data is available for slow jobs.

Further, each head offset address + 8, +
From addresses 16 and 24, each is a 4-word memory,
It is designed to be used in 4ms jobs. In each of these memories, the latest A / D conversion result is counted from the beginning,
The value of the overlap counter OLC of the ms job is stored in the subsequent word, the data is read from the head direction when the 4 ms job is executed, and the data of each subsequent word is sequentially shifted to the head direction as the job ends. Therefore, it is a FIFO buffer that is read out from the data that has been stored previously.

That is, the A / D conversion is accurately performed in a cycle of 4 ms by a periodic interrupt every 0.5 ms.
ms jobs may be delayed by being disturbed by higher priority jobs. Therefore, the FIFO buffer is used for the transfer of the A / D conversion, and +8 to +16 to +24 for a 4 ms job.
After referring to the data in the FIFO buffers at addresses, the data in the FIFO buffers are sequentially shifted in step S526.

On the other hand, when the job to be executed is not the 4 ms job in step S520 but branches to the label ALJ30, the job to be executed is 10 in step S530.
It is checked whether it is an ms job or not, and if it is a 10 ms job, the work area of the job is set in step S531, and the step
If the level 0 interrupt is enabled in S532, step S5
At 33, the section moves to the 10 ms job section, the job main body is executed, and at step S534, interruption is prohibited and the routine exits.

In the 10 ms job section, a service routine for calculating the engine speed from the half rotation time, a service routine for creating the above-mentioned ignition schedule, etc. are prepared on the OS side.

If it is determined in step S530 that the job to be executed is not a 10 ms job, the process branches to step S540 to check whether the job to be executed is a low priority class job. If the job is not the low priority class job, the process branches from step S540 to the label ALJ50. If the job to be executed is the low priority class job, the process proceeds from step S540 to step S541, and the current state is the multiplex waiting state. Check if there is.

When the current state is not the multiplex waiting state, the routine proceeds from step S541 to step S542 where the system variable S_ACAS (crank total position variable) is set as the user variable ACAS and the routine proceeds to step S544.
When in the multiplex waiting state, the process branches from step S541 to step S543, and the user variable ACAS is incremented by 1 to 1
After taking the remainder divided by 2, proceed to step S544.

In step S544, the work area of the job is set, and in step S545, when the interrupt of level 0 is permitted, in step S546, the low priority class job is moved to the section, the job body is executed, and then step S547.
Disable the interrupt and exit the routine.

Further, in the label ALJ50, the step
In S550, it is checked whether or not the job to be executed is a 50 ms job. If the job is a 50 ms job, the process proceeds to step S551 to set the work area of the job, and the process proceeds to step S552.

When the level 0 interrupt is permitted in step S552, the process moves to the 50 ms job section in step S553, and the stalling processing routine prepared on the OS side,
The ignition timing retard routine for each cylinder, the fuel injection start timing calculation routine, etc. are executed, and the routine based on the control strategy on the user side is executed. After the job is completed, the interrupt is prohibited in step S554, and the routine is exited.

On the other hand, if the job to be executed in step S550 is not a 50 ms job, the process branches to step S560 to set the work area of the job, permit the level 0 interrupt in step S561, and proceed to step S562.
After moving to the section area of the 250 ms job and executing the job body, the interrupt is prohibited and the routine exits in step S563.

In the above job priority processing, it is necessary to always know the crank position accurately, and the crank position calculation subroutine shown in FIG. 8 is executed for each classen interrupt to execute the crank position variable S_CCAS,
S_RCAS, S_ACAS, S_ECAS are calculated. In the following description, S_ indicating a system variable will be omitted from the crank position variable.

In this crank position calculation subroutine,
First, in step S600, the lower 2 bytes of the class timer are stored in the lower 2 bytes of the soft timer. This classen timer is a hardware timer provided in the ECU 50, and in the present embodiment, a 16-bit timer has a maximum of 2 timers.
Counting is possible up to 55 ms, but a continuous area of 3 bytes is secured in the memory, 2 bytes of the Classen timer are transferred to the lower 2 bytes, and the 3rd byte is counted up by the interrupt generated by the overflow of the Classen timer. It is used as a soft timer. As a result, the signal input interval time from the crank angle sensor 39 (hereinafter abbreviated as “Krasen interval”) can be up to 64 sec (2
It is possible to count up to 55 ms x 256), and 1
Without using a special hardware timer of 6 bits or more, it is possible to easily cope with an extremely long Krasen interval such as cranking.

Next, when proceeding to step S601, it is checked whether or not the class interval is equal to or shorter than the set time. This set time is
The time as a classen interval corresponding to the maximum engine speed is, for example, 0.3 ms, and when the classen interval is less than or equal to the set time in the above step S601, it is determined as a counting error of the classen timer due to mixing of noise or the like in step S602.
In step S601, the error code 1 is stored in the accumulator A, and the routine is exited. If the classen interval is longer than the set time in step S601, the classen timer count is considered to be normal, and the process proceeds to step S603.

At step S603, CCAS.R to be described later is performed.
The CAS determination subroutine is executed to determine the crank position, and in step S604, it is determined whether or not the error level ECAS is 2, that is, whether or not the cylinder is not determined, such as during cranking. Examine, ECAS
= 2, the process branches to step S605, the error code 1 is stored in the accumulator A, and the routine exits.

On the other hand, in the above step S604, ECAS ≠ 2
In case of, the process proceeds to step S606, and the third byte of the soft timer is set to 0. Then, the process proceeds to step S607, whether or not the crank position information variable CCAS is 1, that is,
Current crank position is BTDC 65 ° CA-10 ° CA
During the period (see FIG. 23), CCAS = 1
In case of, it jumps from step S607 to step S609, and in case of CCAS ≠ 1, it advances from step S607 to step S608 and increments the A / D conversion request by 1,
Proceed to step S609.

This A / D conversion request is a flag-like variable for instructing the crank synchronous A / D conversion for each crank angle 90 ° CA. It takes a value of 0, 1 and when the value is 1, the crank Instructs synchronous A / D conversion. That is, as described above, the A / D conversion of 1 channel out of the 8 channels A / D conversion is performed at every crank angle 90 ° CA, but when CCAS becomes 0 (BTDC 97 ° C.
A) and when CCAS becomes 2 (BTDC 10 ° C
A), A / D conversion request for crank synchronization is set, and A / D conversion is performed at every 90 ° crank angle (accurately, every 87 °, 93 °) for the A / D conversion order every 0.5 ms. Is interrupted.

Thereafter, in step S609, the total position variable ACAS is calculated by multiplying the cylinder information variable RCAS by 3 and adding the crank position information variable CCAS (ACAS = RCAS × 3 + CCAS), and then step S610.
Then, the soft timer is limited by 2 bytes, and if the classen timer overflows, the lower 2 bytes are F
The processing proceeds to step S611 as FFF (255 ms).

In step S611, the total position variable ACAS
Array TCAS [AC with subscripts = 0, 1, 2, ... 11
AS] (element of the array), the Krasen interval data is stored, and in step S612, the crank position information variable CCAS =
Array MTCSX [CCAS] with subscripts 0, 1, 2
Store Classen interval data in (array element).

As shown in FIG. 21A, the array TCAS is a 12-word classen interval table in which classen interval data for two engine revolutions corresponding to ACAS = 0, 1, 2, ... 11 is stored. , The array MTCSX is
As shown in FIG. 6B, it is a 3-word class interval table in which three class interval data corresponding to CCAS = 0, 1, 2 are stored.

That is, the CCAS
RCAS discrimination subroutine (details will be described later), and each information variable CCAS, RCAS, AC by step S609.
AS is updated, for example, CCAS = 1, RCAS =
1, updated to ACAS = 4, and when the crank position is currently BTDC 65 ° to 10 ° CA of the # 3 cylinder, from the classen signal input at the BTDC 97 ° CA of the # 3 cylinder measured by the classen timer to the # 3 cylinder. BTDC
The time until the class signal is input at 65 ° CA (class interval data) is calculated in step S611 as the total position variable AC.
The AS is stored as a parameter at the address ACAS = 3 in the array TCAS, and the crank position information variable CCAS is used as a parameter in the array MTCS in step S612.
Store to address CCAS = 0 of X.

Therefore, the total position variable ACAS and the crank position information variable C are set for each classen interrupt by the classen input.
Every time the CAS is updated, the data in the arrays TCAS and MTCSX are sequentially updated. Therefore, by referring to the array TCAS, it is possible to know the change in the Krasen interval (change in rotational speed) at each crank position of each cylinder. Therefore, it is possible to determine whether or not there is a misfire in each cylinder, the combustion state, and the like, and it is possible to grasp the operating status of all cylinders. Also,
By referring to the array MTCSX, the latest Classen interval can always be obtained, and the current operating condition can be quickly grasped.

Next, in step S613, the value of the error level ECAS is checked again. Here, since it has already been confirmed in step S604 that ECAS ≠ 2, it is checked whether or not the error level ECAS is 1, that is, whether or not the determination of the crank position is an estimation state in which anxiety remains. .

When ECAS ≠ 1 (that is, ECAS = 0) in step S613 and the crank position is determined with certainty, the process proceeds from step S613 to step S614, and the sum of the latest three classen interval data ( The sum of the Classen interval data stored in the array MTCSX) is calculated as a 3-byte half rotation time MTCS18 (MTCS18 = ΣMTCSX). That is,
The half rotation time MTCS18 is calculated each time the crank position information variable CCAS is updated for each classen interrupt and the data in the array MTCSX is updated, and the BTDC 97 °,
The latest data can always be obtained by taking the moving sum of the Krassen intervals for each position of 65 ° and 10 ° CA.

On the other hand, in the above step S613, ECAS = 1
When the crank position is in an estimated state where the determination of the crank position remains uncertain, the half rotation time MTCS1 is calculated from the array MTCSX.
8 is not calculated, but the half rotation time estimation subroutine is executed in step S615 to estimate the half rotation time MTCS18.

The estimation of the half rotation time MTCS 18 is to estimate the current half rotation time from the previous Krasen interval time. That is, as shown in FIG. 23, if the current crank position information variable CCAS is 1, the previous crank angle is 32 ° CA (BTDC 97 ° to 65 ° C).
The angle between A) and the half-rotation time MT based on the following equation based on the crank angle and the previous Krasen distance data.
Calculate CS18. MTCS18 = previous classen interval × 180/32 Also, when CCAS = 2, BT
DC 55 ° CA between 65 ° and 10 ° CA and the previous Krasen interval data, the previous Krasen interval x 180 /
The half rotation time MTCS18 is calculated from 55, and
When CCAS = 0, BTDC 10 ° CA to ATD
Based on the angle 93 ° CA between C83 ° CA (BTDC97 ° CA of the next cylinder) and the previous Krasen interval data, the half-rotation time MTC is calculated by the previous Classen interval x 180/93.
S18 is calculated.

Then, in step S614, the half rotation time M
After calculating the TCS18 or after estimating the half rotation time MTCS18 in the above step S615, the process proceeds to step S616, and the half rotation time MTCS18 of 3 bytes is limited to 2 bytes and stored in a predetermined variable MTCSK. In step S617, this variable MTCSK is doubled and the variable MT
Store it in CSK4, store the normal end code 0 in accumulator A in step S618, and exit the routine.

Then, in the above-mentioned job execution subroutine, the engine speed is calculated every 10 ms, and this engine speed is calculated from the reciprocal of the variable MTCSK which limits the half-rotation time MTCS18 of 3 bytes to 2 bytes.

More specifically, half of the unit time (1 min) of the revolutions per minute rpm, 30 seconds, is taken as the half revolution time MT.
Divide by CSK to make 1 rpm as a unit 2
The variable NRPM in bytes, that is, the engine speed is calculated (NRPM = 30 sec / MTCSK), and this engine speed is used as one of the basic parameters in various control amount calculation processes.

Next, according to the flow chart shown in FIG. 9, step S603 of the above crank position calculation subroutine.
The CCAS / RCAS discrimination subroutine executed in step S1 will be described.

In this subroutine, first, in step S800, the CamSen counter is limited to 0-4. Cam angle sensor 4 counted by this cam sensor counter
As shown in FIG. 23, the number of cam pulses from 1 (the number of cam pulses between Krasen signal inputs) is 0 in the normal state.
Although it is 3, since there is a possibility that an abnormal count value of 4 or more will be caused by the influence of noise, etc.
It is limited to 4 and represents an abnormal state by 4.

Next, in step S801, a combination of 5 × 4 × 2 (5 in which the CamSen counter is 0 to 4 is selected) from (the count value of) the CamSen counter, the cylinder information variable RCAS, and the crank position information variable CCAS. Type, cylinder information variable RCAS
Of the cylinder / crank position state CCHMA in which the state data for four types of 0 to 3 and two types when the crank position information variable CCAS is 0, 1 and 2 are stored.
Read P.

This cylinder / crank position state map CCH
As shown in FIGS. 22 (a) and 22 (b), the MAP is used when the crank position information variable CCAS is 0 or 1, and when the crank position information variable CCAS is a change point of the cylinder information variable RCAS.
The state data indicating whether normal, abnormal, or determinable should be estimated is stored for all possible states of each combination of the cam sen counter and the cylinder information variable RCAS. The current state can be evaluated and the next state to be taken can be known.

This state data is 2-bit data, and the value of bit 0 represents definite or estimated, and the value of bit 1 represents normal or abnormal. The value of bit 0 is
When the value is 0, the determination is made, and when the value is 1, the estimation is shown. As can be seen from FIG. 22, the determination is made only when the Camcen counter is 2 or 3 (see the Camcen signal pattern shown in FIG. 23), and the other values are estimated. It is inevitable. The value of bit 1 is normal when it is 0 and abnormal when it is 1, and is normal only when the camsen counter is 3 or less and the combination of crank position variables shown in FIG. 23 is met. Is an abnormal condition.

For example, CCAS = 0 or 1, that is,
As can be seen from FIG. 23, the combination in which the CAMDC counter is 0 and the cylinder information variable RCAS is 0 for BTDC 97 ° to 10 ° CA of a certain cylinder is in a state in which the crank position can be normally estimated. , The cylinder shown in FIG.
Binary 01 (normally estimated) state data is stored in the corresponding region of the crank position state map CCHMAP. Further, when the CamSen counter is 1 and the cylinder information variable RC
Since the combination in which AS is 0 is obviously abnormal and is a state that can only be estimated, state data of 11 (abnormality estimation) in binary number is stored in the corresponding region of the cylinder / crank position state map CCHMAP. .

Further, CCAS = 2, that is, a combination of BTDC 10 ° CA to ATDC 83 ° CA of a certain cylinder, in which the Camsen counter is 3 and the cylinder information variable RCAS is 0, the cranks including the TDC of the # 1 cylinder are used. Cylinder / crank position status map C because the position can be normally determined
The combination in which the state data of 00 (normally determined) in binary number is stored in the corresponding area of CHMAP, and further, the combination in which the Camcen counter is 2 and the cylinder information variable RCAS is 0 is obviously abnormal, but the Camcen input is Since there are two or more states that must be determined, the state data of 10 (abnormal determination) in binary is stored in the corresponding region of the cylinder / crank position state map CCHMAP.

When the status data is read from the cylinder / crank position status map CCHMAP in step S801, the process proceeds to step S802, and it is determined whether or not the error level ECAS is 2, that is, the current status is not cylinder-determined. Check whether it is in an unknown state, ECAS = 2
If it is, it is checked whether or not bit 0 of the state data read from the cylinder / crank position state map CCHMAP in step S803 is 0, that is, whether it is the confirmed state. If it is the confirmed state, the process proceeds to step S804, and the state is not the confirmed state. When it is in the estimated state, it exits the routine and waits until it becomes the confirmed state.

On the other hand, when ECAS ≠ 2 in step S802, the process proceeds to step S804 to check whether the state is the estimated state, and depending on the confirmed state and the estimated state, the process proceeds to step S805 and subsequent steps or step S812 and subsequent steps. . If the process proceeds to step S804 in the confirmed state at step S803, the process proceeds to step S805 and subsequent steps.

First, the processing from step S805 onward will be described.
Since it is in the confirmed state that the cylinder discrimination is made regardless of abnormality,
As can be seen from the time chart of FIG. 23, the crank position information variable CCAS is set to 0 because the classen interrupt this time is an interrupt of BTDC97 ° CA after three or two cam pulses are input.

Next, in step S806, it is checked whether or not the Camcen counter is 3, and the Camcen counter is 3
If it is not, that is, when the CamSen counter is 2, it is after ignition of the # 2 cylinder, so the cylinder information variable RCAS is set to 3 in Step S807 and the process proceeds to Step S809. Since it is after ignition, the cylinder information variable RCAS is set to 1 in step S808, and the flow proceeds to step S809.

In step S809, the status data is read again from the cylinder / crank position status map CCHMAP using the updated crank position information variable CCAS, cylinder information variable RCAS, and Camsen counter as parameters, and bit 1 of this status data is 1 , That is, whether or not it is in an abnormal state.

As a result, in step S809, when bit 1 of the state data is 1 and it is determined that the state is abnormal, the crank position information variable CCAS and the cylinder information variable RC are determined.
Assuming that the AS update result is an uneasy estimate
In step S810, the error level ECAS is set to 1, and the routine exits. When bit 1 of the status data is 0, indicating a normal state, the error level ECAS is set to 0 in step S811, and the routine exits.

On the other hand, in the processing after step S812, in step S812, the current crank position information variable CCAS (the crank position information variable CCAS calculated by the previous Krasen interrupt) is 2, that is, the change point of the cylinder information variable RCAS. If CCAS = 2, the process proceeds from step S812 to step S813, and the cylinder information variable R
CAS is incremented by 1, the crank position information variable CCAS is set to 0 in step S814, and the flow proceeds to step S817.

On the other hand, in the above step S812, CCAS ≠ 2
If it is, the process proceeds from step S812 to step S815 to check whether the Camcen counter is not 0. If the Camcen counter is not 0, the process branches to the above-mentioned step S813, and when the Camcen counter is 0, the process proceeds to step S816.
Then, the crank position information variable CCAS is incremented by 1, and the process proceeds to step S817.

In step S817, the status data is read again from the cylinder / crank position status map CCHMAP using the updated crank position information variable CCAS, cylinder information variable RCAS, and Camsen counter as parameters to determine whether or not there is an abnormal condition. Check, bit 1 of status data is 0
In the normal state, the routine is exited as it is (the current error level ECAS = 0 remains). If the bit 1 of the status data is 1 and the status is abnormal in step S817, the process proceeds to step S818, and the error level ECAS is set as the estimation result that the crank position information variable CCAS and the cylinder information variable RCAS are anxious. Set to 1 and exit the routine.

As described above, the crank position variable CCA
It is possible to grasp the state of the crank position that changes momentarily by S, RCAS, ACAS, and ECAS, and further evaluate the current state by using the cylinder / crank position state map CCHMAP to know the state to be taken next. Therefore, the program can be greatly simplified compared to the processing based on the normal condition judgment, the memory capacity can be saved, the processing speed can be improved, and the flexibility of the program can be improved to change the specifications. Can be dealt with flexibly.

Next, the fuel injection start timing calculation routine shown in FIGS. 10 and 11 will be described. This fuel injection start timing calculation routine is one of the 50 ms jobs executed in step S553 of the job execution subroutine described above, and sets the injection start section, the number of injections, and the injection width per injection for all cylinders. . That is, when the start timing of continuous injection is fixed when the injection end is fixed, the number of injections is set to two, while when the start timing of continuous injection is early, the number of injections is set to one.

First, in step S1000, it is determined whether or not the engine is starting. If the engine is starting, the process proceeds to step S1001, the injection start delay time IJDELY is set to 0, and the effective injection width to be injected per injection is set to 1 in step S1002. Required injection width IJSEIn (n = # 1, # 2, #, which is the total injection amount per cycle)
3, # 4) is set to 1/2, the injection start section setting flag IJTMGF is set to a plurality of injections of 128 + 2, for example, in step S1003, and the routine exits.

The injection start delay time IJDELY is incorporated in the timer waiting time set by the injection timer set macro 2 described later, and the start of the injection start section (BTDC
If it is a single injection after several ms from the input of 97 ° CA class signal (hereinafter abbreviated as “97 ° CA class”) or BTDC 10 ° CA class signal (hereinafter abbreviated as “10 ° CA class”) If the first injection is the second injection, it is determined whether to start the second injection.

Further, the injection start section setting flag IJT is set.
The MGF specifies the injection start section of the # 1 to # 4 cylinders. That is, this injection start section is the injection start section table IJnTBL (n = # 1, #) for each of # 1 to # 4 cylinders.
2, # 3, # 4).

As shown in FIG. 23, each injection start section table IJnTBL (n = # 1, # 2, # 3, # 4).
Then, one cycle of the # 1 to # 4 cylinders corresponds to the above-mentioned crank total position variable ACAS and 97 ° CA class and 1
The injection start section is set by the injection start section setting flag IJTMGF as a 1-byte variable corresponding to each of bits 7 to 0 by dividing the section into 8 sections with 0 ° CA class. That is, in the present embodiment, each injection start section of the injection start section table IJnTBL has 0 in the latter half of the intake stroke, 1 in the first half, 2 in the second half of the exhaust stroke, 3 in the first half, and 4 in the second half of the combustion stroke. 5, the latter half of the compression stroke is defined as 6, and the first half, that is, immediately after the end of the intake stroke is defined as 7. However, the injection process is not performed in the injection start section 0.

Therefore, in step S1003, I
JTMGF = 128 + 2 is 1 if expressed by an 8-bit variable
0000010, the first injection is started from the injection start section 7 (10 ° CA class), and the second injection is started from the injection start section 1 (10 ° CA class) without waiting time. It

On the other hand, if it is determined in step S1000 that it is not the time to start and the process branches to step S1004, it is determined whether the injection start set time IJTIME is zero, and IJTIM is set.
If E = 0, the process returns to step S1001 and step S100.
The routine exits through step S1002, step S1003.

The injection start set time IJTIME is set by the control strategy on the user side.
This injection start set time IJTIME requires the required injection width IJ
By substituting SEIn (n = # 1, # 2, # 3, # 4), the injection end can be almost fixed to BTDC 10 ° CA just before the intake stroke. Further, the injection end control can be performed by arbitrarily setting the injection start set time IJTIME. Therefore, it may be possible to set IJTIME = 0 in the control strategy on the user side. In this case, the process returns from step S1004 to step S1001 and the injection width per injection with respect to the required injection width in step S1002 is set. After setting, a plurality of injection start sections are set in step S1003, and the routine exits.

In step S1004, IJTIME ≠ 0.
If it is determined that the operation proceeds to step S1005, this step S10
In 05 or less, the injection start section and the injection start delay time IJDELY based on the injection start set time IJTIME are calculated.

That is, the injection start set time IJT
The IME is set to the classen interval for each injection start section (BTDC97
° -10 ° CA or BTDC10 ° -ATDC83 ° C
The time interval between A) is sequentially subtracted until it becomes zero or negative to identify each injection start section of multiple injections, and the start of the last injection start section (97 ° CA class or 10
The injection start delay time IJDELY sets the number of ms after which the fuel injection should be started.

First, in step S1005, the Krasen interval is subtracted from the injection start set time IJTIME and the value is stored in the accumulator A. In step S1006, it is determined whether the value of the accumulator A is positive, zero or negative. , A ≦ 0, the routine proceeds to step S1007, where the injection start delay time IJDELY is set by the 2's complement of the value stored in the accumulator A (IJDELY = −A), and at step S1008 the effective injection to be performed once The width is 1 / of the required injection width IJSEIn (n = # 1, # 2, # 3, # 4)
2, the injection start section setting flag IJTMGF is set to 128 + 4 multiple injections in step S1009, and the routine exits. IJTMGF = 128 + 4 = 8
Expressed in a bit variable, it is 10,000100, and therefore the injection start sections are 7 and 2, the first injection is started from the injection start section 7 without waiting time, and the subsequent injection is 97 of the injection start section 2. It is started after the injection start delay time IJDELY from CA Classen.

On the other hand, if it is determined that A> 0 in step S1006, the process branches to step S1010, and the value of the accumulator A is set to a value obtained by subtracting the Krasen interval from the value stored in the accumulator A (A = A-classen interval), the value of this accumulator A is referred to in step S1011. If A ≦ 0, the process proceeds to step S1012, and the injection start delay time IJDELY is set to the two's complement of the value stored in the accumulator A ( IJDELY = -A), the effective injection width to be injected per injection is set to 1/2 of the required injection width IJSEIn in step S1013, and the injection start section setting flag IJTMGF is set to 128+ in step S1014.
Set to multiple injections of 8 and exit the routine. IJT
If MGF = 128 + 8 is expressed by an 8-bit variable, 100
01000, therefore, the injection start sections are 7 and 3, the first injection is started from the injection start section 7 without waiting time, and the subsequent injection is delayed from the 10 ° CA class in the injection start section 3 from the injection start delay. Started after time IJDELY.

On the other hand, if it is judged at step S1011 that A> 0, the process branches to step S1015, and the value of the accumulator A is set to the value obtained by subtracting the Krasen interval from the value stored in the accumulator A (A = A -Classen interval),
In step S1016, refer to the value of this accumulator A,
If A ≦ 0, the process proceeds to step S1017, the injection start delay time IJDELY is set by the complement of 2 of the value of the accumulator A (IJDELY = −A), and 1 is set in step S1018.
The required injection width IJSEI is the effective injection width for each injection.
It is set to 1/2 of n, the injection start section setting flag IJTMGF is set to a plurality of injections of 128 + 16, for example, in step S1019, and the routine exits. IJTMGF = 12
If 8 + 16 is represented by an 8-bit variable, 10010000
Therefore, the injection start sections are 7 and 4, and the first injection is started from the injection start section 7 without waiting time,
The subsequent injection is started after the injection start delay time IJDELY from the 97 ° CA class in the injection start section 4.

On the other hand, if it is determined in step S1016 that A> 0, the process branches to step S1020, and the value of the accumulator A is set to a value obtained by subtracting the Krasen interval from the value stored in the accumulator A (A = A -Classen interval),
The value stored in the accumulator A is referred to in step S1021, and if A ≦ 0, the process proceeds to step S1022, and the injection start delay time IJDELY is set by the complement of 2 of the value of the accumulator A (IJDELY = −A), Step S1
In 023, the effective injection width to be injected per injection is the required injection width I
It is set to 1/2 of JSEIn, the injection start section setting flag IJTMGF is set to, for example, 128 + 32 multiple injections in step S1024, and the routine exits. IJTMG
If F = 128 + 32 is represented by an 8-bit variable, 1010
0000, therefore, the injection start sections are 7 and 5, the first injection is started from the injection start section 7 without waiting time, and the subsequent injection is delayed from the injection start section 5 of 10 ° CA class and the injection start delay. Started after time IJDELY.

On the other hand, if it is determined in step S1021 that A> 0, the process branches to step S1025, and the value of the accumulator A is set to a value obtained by subtracting the Krasen interval from the value stored in the accumulator A (A = A -Classen interval),
The value stored in the accumulator A is referred to in step S1026. If A ≦ 0, the process proceeds to step S1027, and the injection start delay time IJDELY is set to the 2's complement of the value of the accumulator A (IJDELY = −A), Step S1
In 028, set to jet all at once, and step S1029
Then, the injection start section setting flag IJTMGF is set to a single injection of 64, for example, and the routine is exited. IJTMGF
= 64 with an 8-bit variable is 01000000, and the fuel injection is started after the injection start delay time IJDELY from the 97 ° CA class in the injection start section 6. On the other hand, if it is determined that A> 0 in step S1026, the process branches to step S1030, and the value of the accumulator A is set as a value obtained by subtracting the Krasen interval from the value stored in the accumulator A (A = A-classen interval). ), Step S1031
And refer to the value stored in this accumulator A, and A ≤ 0
In the case of, the process proceeds to step S1032, and the injection start delay time IJ
DELY is set by the two's complement of the value stored in the accumulator A (IJDELY = -A), all injection is set at one time in step S1033, the injection start section setting flag IJTMGF is set to 128, and the routine exits. I
If JTMGF = 128 is represented by an 8-bit variable, 100
The fuel injection is started after the injection start delay time IJDELY from the 10 ° CA class in the injection start section 7.

On the other hand, if it is determined in step S1031 that A> 0, the process branches to step S1035, and the injection start delay time I
JDELY is set to 0, and it is set in step S1036 that all injection is to be performed once. In step S1037, the injection start section setting flag IJTMGF is set to 128, and the routine exits.
If IJTMGF = 128 is represented by an 8-bit variable, 10
The fuel injection is started from the 10 ° CA class in the injection start section 7 without waiting time. Therefore, the injection in this step becomes the maximum injection amount in one cycle.

FIG. 24 shows a time chart of an example of setting the fuel injection start timing based on the above flow chart.

As shown in FIG. 13A, in the case of continuous injection, if the injection can be completed in one cycle even if the injection start timing is delayed, the injection is performed twice, and the fuel is injected in the first injection. In order to accelerate the vaporization of the air-fuel ratio, the air-fuel ratio is totally adjusted by the subsequent injection. For example, the injection start section setting flag IJ
TMGF is set to 128 + 8, injection start delay time I
When JDELY is set to the time t, first, from the 10 ° CA class in the injection start section 7, ½ of the total injection amount is injected without waiting time, and then 10 ° in the injection start section 3.
The remaining 1/2 is injected after a time t from CA Classen.

Further, as shown in FIG. 13B, in the case of continuous injection, if the injection cannot be completed in one cycle unless the injection is started early, the injection is performed once. For example, when the injection start section setting flag IJTMGF is set to 64 and the injection start delay time IJDELY is set to the time t, first, all the injections are made after the time t from the 97 ° CA class in the injection start section 6.

In the above flowchart, the bit designation of the injection start section setting flag IJTMGF is one or two per cylinder, but it is also possible to designate three or more bits by the injection start section setting flag IJTMGF. . By making multiple injections, the air-fuel ratio is totally adjusted at the last injection, and as a result, it is possible to flexibly respond to the fluctuation of the air-fuel ratio during transition such as acceleration increase by adjusting the subsequent injection amount. be able to.

Further, in this flowchart, since the injection start section is set every 50 msec, even if the engine speed reaches a high speed such as 9000 to 10000 rpm, the load on the main CPU 58 does not increase, and the high speed is maintained. The injection end control becomes possible even in the rotational speed range.

FIG. 12 is an injection timer setting subroutine, which is executed in step S205 of the interrupt processing routine by the classen shown in FIG. 2, and sets the injection timer for each injection start section.

First, in step S1100, the error level EC
If AS is 2, and ECAS = 2, the crank position is still unknown, so the subroutine is skipped without further processing. If ECAS ≠ 2, the step is executed.
The process proceeds to S1101, and it is determined whether the crank position information variable CCAS is 1, and when CCAS = 1, it is determined that calculation is unnecessary and the routine exits. As shown in FIG. 23, in the case of CCAS = 1, it is an interruption of 65 ° CA class, and since the setting of the injection timer in the current injection start section has already been completed, the subroutine is immediately exited and the processing load is reduced. On the other hand, if CCAS ≠ 1, it is either a 97 ° CA class interrupt or a 10 ° CA class interrupt, which is the start of the injection start section, and therefore injection in the current injection start section in step S1102 and subsequent steps. Set the width and injection start timing.

First, in steps S1102 to S1105, the injection timer set macro 1 described below is executed for all cylinders, and the injection width (hereinafter referred to as "effective injection width") TEIJAn (n = # 1, in the current injection start section). # 2, #
3, 3) are set for all cylinders.

Then, in step S1106, the effective injection width TE set in each of the above steps S1102 to S1105.
It is determined whether IJAn is all zero. If all IJAn are zero, it is not necessary to reset the injection timer and the process exits the subroutine. Further, at least one cylinder has TEIJAN ≠ 0.
In this case, the process proceeds to step S1107, all the injection timers are stopped, and during that time, the processes of steps S1108 to S1111 described below are performed.

In steps S1108 to S1111,
The injection timer set macro 2 described below is executed for all cylinders, and the injection time based on the effective injection width TEIJAn set by the injection timer set macro 1 is set for all cylinders.

Then, in step S1112, all the injection timers are restarted and the subroutine is exited.

In step S1107, when all the injection timers are stopped, the injection time is not measured, so that the fuel injection is continued in the fuel-injected cylinders and the injection stopped state is continued in the injection-stopped cylinders. To be done. The processing time corresponding to the time when the injection timer is stopped is taken into consideration in the injection time set in steps S1108 to S1112.

The injection timer set macro 1 for each cylinder executed in steps S1102 to S1105 in the above-mentioned injection timer set subroutine is executed according to the flowcharts of the injection timer set macro 1 shown in FIGS. The program of the injection timer set macro 1 is a macro program so that the injection timer set routine of the # 1 cylinder is used as a basic program and can be applied to the # 2 to # 4 cylinders.

First, at step S1200, the corresponding #n (n =
1, 2, 3 or 4) the current injection start section number of the cylinder,
Based on the crank total position variable ACAS, the injection start section table IJnTBL (FIG. 23) of the corresponding #n cylinder is referred to, and it is determined in step S1201 whether the injection start section number is zero. If the injection start section is zero, the process proceeds to step S1202, and if the injection start section is not zero, the process proceeds to step S1210.

In the injection start section 0, the injection process is not performed since the intake stroke is in the latter half. Therefore, in steps S1202 to S1209, the fuel injection initialization process is executed in preparation for the next fuel injection. First, in step S1202, the cumulative injection width TEIJBn of the #n cylinder is set to zero, and step S1203
Determines whether there is a fuel cut request from the job on the user side, and if there is no fuel cut request, the process proceeds to step S1204,
If there is a fuel cut request, the process advances to step S1208.

When it is judged that there is no fuel cut request, step S1
When the process proceeds to 204, it is determined whether the #n cylinder is in the fuel cut state by referring to the value of the fuel cut state flag PRCUT. If it is not in the fuel cut state, the process proceeds to step S1205, and if it is in the fuel cut state, the process proceeds to step S1206. .

This fuel cut state flag PRCUT is #
1 to # 4 cylinders are shown in correspondence with 4 bits of bit 0 to bit 3, and are set in the first routine when a fuel cut is requested in the user side job, and the first after the fuel cut request is released. It is cleared in the routine.

As described above, the fuel cut state flag PRCUT determines whether or not the bit corresponding to the #n cylinder is set in the injection start section 0, there is no fuel cut request from the user side job, and Fuel is not injected into the cylinder for which the fuel cut state flag PRCUT is set in the next cycle.

If it is determined that the fuel cut state is not reached and the routine proceeds to step S1205, the injection process is not performed in the injection start section 0, so the effective injection width TEIJ of the #n cylinder concerned.
An is set to zero, and the injection timer set macro 1 for the #n cylinder is ended.

If it is determined in step S1204 that the fuel cut state has been reached and the process proceeds to step S1206, there is no fuel cut request from the user side job and the fuel cut state flag PRCUT is set. In order to prepare for the fuel cut recovery in the next cycle, the fuel corresponding to the adhesion correction amount is injected once at the injection start section zero of the #n cylinder. Therefore, the fuel adhesion correction amount TEFUCH set by the user side job is set. In step S1207, the effective injection width TEIJAn of the #n cylinder is updated.
After clearing the bit corresponding to the #n cylinder of the fuel cut state flag PRCUT, the injection timer set macro 1 of the #n cylinder is ended.

As a result, it is possible to set the adhesion correction for the fuel cut and the fuel cut recovery for each cylinder.

If it is determined in step S1203 that there is a fuel cut request from the user side job, it is effective in step S1208 because the fuel cut is requested in the injection start section 0 in which the injection process is not performed. Injection width TE
IJAn is cleared, the bit corresponding to the #n cylinder of the fuel cut state flag PRCUT is set in step S1209, and the injection timer set macro 1 of the #n cylinder is ended.

If it is determined in step S1201 that the current injection start section is other than zero, that is, the section of the injection processing region, normal injection processing is performed in step S1210 and thereafter. First, in step S1210, the bit value of the #n cylinder of the fuel cut state flag PRCUT is referenced to determine whether the #n cylinder is in the fuel cut state. If it is in the fuel cut state, the process proceeds to step S1211 and the effective injection width TEIJAn is set to zero, and the injection timer set macro 1 of the #n cylinder is ended.

[0187] On the other hand, if it is determined in step S1210 that the fuel cut state is not established, then in step S1212 it is determined whether or not injection will be performed in the section after this time.
If it is determined with reference to the value of TMGF, and it is determined that the injection will be performed in the section after this time, the flow proceeds to step S1213,
Also, if it is determined that injection will not be performed in the section after this,
Proceeds to step S1218.

Step S1212 is a step corresponding to a plurality of injections in one cycle, and when there are two or more sections designated by bits 0 to 7 of the injection start section setting flag IJTMGF, for example, as shown in FIG. ), If IJTMGF = 128 + 8, sections 7 and 3 are injection start sections, and in the injection start section 7, injection is performed in a section after this, so the flow proceeds to step S1213, while the injection start section In 3, the injection is not started in the section after this, so the flow proceeds to step S1218.

When proceeding to step S1213, it is determined whether the injection is to be performed in the current section, the injection start section setting flag IJ.
Referring to the value of TMGF, the indicated value is set in the above step S120.
If it is judged that the injection number does not coincide with the current injection start section number searched in 0, and if the injection is not performed in the current section, step S121
4, the effective injection width TEIJAn is set to zero and the relevant #n
The cylinder injection timer set macro 1 is terminated.

On the other hand, if it is determined in step S1213 that the injection is to be performed in the current section, the flow advances to step S1215 to divide the cylinder-specific required injection width IJSEIn set every 10 ms by the user's job by 2 n. Then, the injection amount per time is calculated, the value is stored in the accumulator A, the effective injection width TEIJAn of the #n cylinder is set by the value stored in the accumulator A in step S1216, and the accumulator is set in step S1217. Cumulative injection width T to the value of A
EIJBn is added and the injection timer set macro 1 ends.

The above order n is determined by the steps S1002, S1008, S1013, S1018, S1 in the above-described fuel injection start timing calculation routine.
It is set based on the number of injections set in 023, S1028, S1033, or S1036, and n = 1 when 1/2 injection is set and n when all injections are set.
Is set to = 0. As a result, in the double injection, the latest required injection width IJSE for each cylinder is obtained in the last injection start section.
An effective injection width corresponding to 1/2 of In is set.

In addition, the above steps S1212 to S12
When it proceeds to step 18, it is determined whether to inject in the current section in the above step S1.
Similar to 213, the injection start section setting flag IJTMGF
If it is determined that the injection is performed in the current section, the process proceeds to step S1219. If the injection is not performed in the current section, the process proceeds to step S1223.

Then, in step S1219, a value obtained by subtracting the cumulative injection width TEIJBn from the cylinder-specific required injection width IJSEIn is stored in the accumulator A, and step S122
At 0, the effective injection width TEIJ at the value of this accumulator A
Set An.

Next, at step S1221, the cumulative injection width TE
Step S1 with IJBn as the required injection width IJSEIn
Continue to 222. That is, since the current section is the injection end section and the injection is not performed in the section after this, the cumulative injection width TEIJBn is set to the required injection width IJ in this routine.
Set to the same value as SEIn. As a result, for example, in the case of injection once in one cycle, TEIJBn = 0, so TEIJAn = IJSEIn. Then, when the routine is executed in the next injection start section, TEIJBn =
IJSEIn = 0.

Then, in step S1222, the injection waiting setting flag FLG_IJ is set and the injection timer set macro 1 is ended. That is, in this routine, it is the last injection in the cycle of the #n cylinder, and by setting the injection waiting setting flag FLG_IJ, the injection end timing set by the injection timer set macro 2 described later is set to, for example, immediately before the intake stroke. Can be adjusted to 10 ° CA Class.

On the other hand, the above steps S1218 to S12
When proceeding to 23, a value obtained by subtracting the cumulative injection width TEIJBn set in step S1221 from the cylinder-specific required injection width IJSEIn is stored in the accumulator A.

Then, in step S1224, the value of the accumulator A is referred to. If A <0, that is, if the operating condition changes and the required injection amount IJSEIn decreases, the process proceeds to step S1226, and A ≧ 0, that is, If the required injection amount IJSEIn has increased or has not changed, the process proceeds to step S1225. Since the fuel injection width is set every 10 ms in the job on the user side, if the requested injection width IJSEIn read at the time of execution of the next routine is different from the requested injection width IJSEIn read at the time of execution of the previous routine, the difference between them. Since the effective injection width TEIJAn can be always set to the latest value by increasing or decreasing the value, the followability to changes in operating conditions is improved.

Thereafter, the above steps S1224 to S
Proceeding to 1225, the value of the accumulator A is compared with the allowable fluctuation value of the injection width, for example, 0.5 ms, and A> 0.5 m
If s, the process proceeds to step S1226. If A ≦ 0.5 ms, the process proceeds to step S1214, and the effective injection width TEIJA.
The n is set to zero and the injection timer set macro 1 of the #n cylinder concerned is ended.

Further, when the process proceeds from step S1224 or step S1225 to step S1226, the effective injection width TEIJAn is set by the value of the accumulator A (negative value from step S1224), and the cumulative injection width TE is set at step S1227.
The injection timer set macro 1 is ended with IJBn being the required injection width IJSEIn.

The injection timer set macro 2 in steps S1108 to S1111 in the above-mentioned injection timer set subroutine is executed according to the flowchart of the injection timer set macro 2 shown in FIGS. The program of this injection timer set macro 2 is based on the timer set routine of the # 1 cylinder as the basic program as in the flow chart showing the injection timer set macro 1, and the macro can be applied to the # 2 to # 4 cylinders. It has been transformed.

First, in step S1300, the injection state of the fuel set in the previous injection start section is judged from the state of the injection timer, and if the injection is continued in this injection start section, the process proceeds to step S1301 and waits for injection. If the state is continued, the process proceeds to step S1307, and if the injection is finished, the process proceeds to step S1312.

[0202] It is determined that the injection is continuously performed, and the step S130 is performed.
If you proceed to 1, the current timer value of the injection timer that is stopped,
The effective injection width TEIJAn (ms) of the positive or negative value set by the injection timer set macro 1 of the #n cylinder is added, and the value is stored in the accumulator A. Therefore,
The value of the injection timer once set in the previous routine is corrected by the latest effective injection width TEIJAn set based on the change in the operating conditions thereafter.

Then, in step S1302, the processing time (the time from the stop of the injection timer to the resetting) is subtracted from the value of the accumulator A, and the subtracted value is used in step S1302.
In S1303, the maximum injection amount is restricted from zero to two rotation times in one cycle (0 ≦ A <2 rotation times, therefore, A = 0 when A is negative), and step S1304
Since the injection is currently in progress, the timer waiting time is set to 0, the reload value of the injection timer is set by the value of the accumulator A in step S1305, and the output pattern of the injection pulse corresponding to the reload value is set in the timer in step S1306. After setting, the injection timer set macro 2 for the #n cylinder is ended.

As a result, the effective injection width TEI set in the present section is obtained in the section after the injection section has passed in the single injection and in the section after the last injection section has been passed in the multiple injections.
When JAn is increased by 0.5 ms or more compared to the value set in the previous section, the injection timer is extended, and when the effective injection width TEIJAn is decreased, injection in the last injection section is performed. The timer is shortened to adjust the total injection amount supplied to the cylinder during one intake stroke.

On the other hand, when it is judged in step S1300 that the injection waiting state is continued and the routine proceeds to step S1307, the timer reload value set in the injection timer is set to the positive or negative value set in this section. Value of effective injection width T
EIJAn is added, and the value is stored in the accumulator A. In step S1308, the value of the accumulator A is limited to two rotation times, which is the maximum injection amount of one cycle from zero (0 ≦ A <2 rotation time). ), Step S1309
In step S1310, the timer reload value is set to the value stored in the accumulator A, and the output pattern of the injection pulse corresponding to the timer reload value is set in step S1311. To end the injection timer set macro 2 for the #n cylinder.

The timer waiting time is set in step S1325 described later in order to adjust the injection end time, but the waiting time set in the previous injection start section is also set in the current injection start section. If the time is still being counted, the timer waiting time is abnormally long, so it is set to zero in step S1309 above, injection is immediately started in the current injection start section, and the injection amount supplied during one intake stroke is adjusted. .

Further, when it is judged in the above step S1300 that the injection in the previous injection start section has already ended, and the routine proceeds to step S1312, the effective injection width TEI is set in the accumulator A.
JAN is stored, and it is determined in step S1313 whether the value of the accumulator A is positive, negative, or zero. If A ≦ 0, the process proceeds to step S1314. If A> 0, the process proceeds to step S1313.
Branch to 18.

In step S1314, the effective injection width TEIJ
Since An has not increased, the injection waiting setting flag FLG_IJ is cleared to end the injection in one cycle, the timer waiting time is set to zero in step S1315, the reload value of the injection timer is set to zero in step S1316, and step S1317. Then, the output pattern is set to zero and the injection timer set macro 2 for the #n cylinder concerned is ended.

[0209] In addition, in step S1313, the effective injection width T
When EIJAn is positive, in step S1318, the dead time TSIN for each cylinder set by another control strategy is set in the accumulator A in order to compensate for the increase in the injection amount.
If Jn (n = # 1, # 2, # 3, # 4) is added, and the value stored in the accumulator A in step S1319 exceeds the maximum injection amount of one cycle, that is, two rotation times, When it is not in the injection waiting state, that is, when the injection waiting state is set, that is, the injection waiting state is determined by referring to the value of the injection waiting setting flag FLG_IJ in step S1320. Flag FLG
If _IJ is cleared, the process proceeds to step S1321, the timer waiting time is set to 0 in order to immediately inject the increased amount, the timer reload value is set at the value of the accumulator A (TEIJAn + TSINJn) in step S1322, and step S1323 is set. Then, the output pattern of the injection pulse corresponding to the timer reload value is set in the injection timer, and the injection timer set macro 2 for the #n cylinder concerned is ended.

On the other hand, if it is determined in step S1320 that the injection is waiting and the process proceeds to step S1324, after the injection waiting setting flag FLG_IJ is cleared, in step S1325,
The timer waiting time is set by the injection start delay time IJDELY set in the injection start timing calculation routine, and the step
The timer reload value is set by the value of the accumulator A in S1326, the output pattern of the injection pulse corresponding to the timer reload value is set in the injection timer in step S1327, and the injection timer set macro 2 for the #n cylinder is ended. . As a result, fuel injection is started after a predetermined timer waiting time (IJDELY) from the start of this injection start section.

As a result of the above, as shown in FIG. 25, the number of injections in one cycle is set to 4, 3, or 2 by the bit designation by the injection start section setting flag IJTMGF set in the calculation routine of the fuel injection start timing. Etc. can be set multiple times. Also, in multiple injections, the effective injection width other than the final injection is set to ¼ of the required injection width by designating the order n.
It can be set arbitrarily such as 1/1, and can be adjusted so that the sum of the injection widths in one cycle in the final injection becomes the latest required injection width. Furthermore, if the injection does not end within the injection start zone of the final injection, the injection timer is extended, and if the effective injection width decreases, the injection timer is shortened to follow the transition during acceleration and other increases. Improve

FIG. 25 (b) shows four injections (IJT
MGF = 128 + 32 + 8 + 2), the effective injection width per injection is 1/4 of the required injection width, and FIG. 7C shows three times injection (IJTMGF = 32 + 8 + 2) effective injection width per injection is required. 1/4 of the injection width, 2 in FIG.
In the single injection (set to IJTMGF = 16 + 2), the effective injection width per injection is 1/1 of the required injection width, and FIG.
It shows an example in which the effective injection width per injection is set to 1/1 of the required injection width in each of two injections (IJTMGF = 4 + 2).

Although one embodiment of the present invention has been described above, the present invention is not limited to this, as long as it is a multi-cylinder engine,
It can also be applied to engines other than horizontally opposed engines.

[0214]

As described above, according to the present invention, one cycle of each cylinder of a multi-cylinder engine is divided into eight sections,
Each of the sections is assigned to bits 0 to 7, and in the interrupt processing synchronized with the engine rotation, it is determined for each cylinder whether the current section is the injection start section based on the instructed bit at one or two or more locations. Therefore, it is easy to specify the injection start section in a plurality of injections, the burden of interrupt processing is reduced, and good fuel injection controllability can be obtained.

[Brief description of drawings]

FIG. 1 is a flowchart showing a periodic interrupt process every 0.5 ms.

FIG. 2 is a flowchart showing Krasen interrupt processing.

FIG. 3 is a flowchart showing job priority processing.

FIG. 4 is a flowchart showing a job execution subroutine.

[FIG. 5] Same as above

[FIG. 6] Same as above

[FIG. 7] Same as above

FIG. 8 is a flowchart showing a crank position calculation subroutine.

FIG. 9 is a flowchart showing a CCAS / RCAS discrimination subroutine.

FIG. 10 is a flowchart showing a routine for calculating a fuel injection start timing.

[FIG. 11] Same as above

FIG. 12 is a flowchart showing an injection timer setting subroutine.

FIG. 13 is a flowchart showing an injection timer set macro 1 routine.

FIG. 14 Same as above

FIG. 15 is a flowchart showing an injection timer set macro 2 routine.

FIG. 16 Same as above

FIG. 17 is an explanatory diagram showing a job execution state.

FIG. 18 is an explanatory diagram of a job flag.

FIG. 19 is an explanatory diagram showing changes in the job execution flag and the overlap counter.

FIG. 20 is an explanatory diagram of a system shift buffer.

FIG. 21 is an explanatory diagram of a classen interval table.

FIG. 22 is an explanatory diagram of a cylinder / crank position state map.

FIG. 23 is a crank position, a cam position, a crank position variable, an injection start section table, and a time chart of stroke order for each cylinder.

FIG. 24 is an explanatory diagram of an injection start section and an injection start section setting flag.

FIG. 25 is a time chart showing injection timer control.

FIG. 26 is a schematic configuration diagram of an engine system.

FIG. 27 is a front view of a crank rotor and a crank angle sensor.

FIG. 28 is a front view of a cam rotor and a cam angle sensor.

FIG. 29 is a circuit configuration diagram of an electronic control system.

[Explanation of symbols]

 1 ... Engine 25 ... Injector 39 ... Crank angle sensor 41 ... Cam angle sensor 50 ... Electronic control unit ACAS ... Crank overall position variable IJnTBL ... Injection start section table IJTMGF ... Injection start section setting flag TEIJAn ... Effective injection width

Claims (1)

[Claims] 1. One cycle of each cylinder of a multi-cylinder engine is divided into eight sections, namely the latter half of the intake stroke, the first half of the intake stroke, the latter half of the exhaust stroke, the first half of the exhaust stroke, the latter half of the combustion stroke, the first half of the combustion stroke, the latter half of the compression stroke, and the first half of the compression stroke. It is divided, and each of these sections is used as an injection start section and bit 0 to bit 7
The method for setting an injection start section for each cylinder of a multi-cylinder engine is characterized in that the injection start section for each cylinder is set on the basis of one or two or more designated bits.