JP2001012267A - Valve system having solenoid drive valve - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce a collision sound generated at the seating of a valve by performing control like delaying a seating speed of the valve. SOLUTION: This valve system, having a first spring energizing a valve of an internal combustion engine in a valve opening direction and a second spring energizing the valve in a closing direction to drive the valve by electromagnetic force, has a first electromagnetic force generating source attracting the valve in an opening direction in accordance with a control current, second electromagnetic force generating source attracting the valve in a closing direction in accordance with the control current, calculation part 55 calculating an operating speed of the valve before its seating, and a control part 56 controlling the control current supplied to the first/second electromagnetic force generating source in accordance with displacement of a piston provided in the internal combustion engine. The control part 56 adjusts switching timing, that is, trigger level Vn+1 of the control current supplied to the electromagnetic force generating source in a side attracting the valve by feedback control based on the operating speed υn calculated by the calculation part 55.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a valve operating apparatus for displacing an intake / exhaust valve of an internal combustion engine by the action of an electromagnetic force, and more particularly to controlling the seating speed of such an electromagnetically driven valve.

[0002]

2. Description of the Related Art In internal combustion engines such as engines, conventionally,
2. Description of the Related Art A valve train that uses electromagnetic force to open and close valves (an intake valve and an exhaust valve) that has been mechanically performed using a timing belt, a cam, and the like has been receiving attention. In this type of valve train, the magnetic attraction (that is, the current supplied to the electromagnetic coil) is changed at a timing corresponding to the movement of the piston in the engine. The important thing is that
The actuation speed of the valve immediately before the valve is seated on the fully closed side or the fully open side, that is, the "seating speed" is sufficiently reduced. When the seating speed of the valve is high, a loud collision sound (hitting sound) is generated at the time of sitting. This tapping sound is generated not only by the collision energy when the valve collides with the valve seat, but also by the collision energy when the armature collides with the stator core. Therefore, in the control of driving the valve by the action of the electromagnetic force, it is important to appropriately adjust the electromagnetic force so that the valve is opened and closed and the seating speed is sufficiently reduced.

[0003]

SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to reduce a collision sound generated when a valve is seated by performing control such that the seating speed of the valve is reduced.

[0004]

According to a first aspect of the present invention, a first spring for urging a valve of an internal combustion engine in a valve opening direction and a valve for urging the valve in a valve closing direction are provided. And a second spring for driving the valve by an electromagnetic force, wherein the first electromagnetic force generating source generates a magnetic attraction force for displacing the valve in the valve opening direction in accordance with a control current. A second electromagnetic force generating source for generating a magnetic attraction force for displacing the valve in the valve closing direction in accordance with the control current; a calculating means for calculating an operating speed of the valve before the valve is seated; And control means for controlling a control current supplied to the electromagnetic force generation source and the second electromagnetic force generation source. Then, the control means adjusts the control current supplied to the electromagnetic force generation source on the side that sucks the valve so that the operation speed calculated by the calculation means becomes slow.

According to a second aspect of the present invention, there is provided a first spring for urging a valve included in an internal combustion engine in a valve opening direction, and a second spring for urging the valve in a valve closing direction. In a valve operating device driven by an electromagnetic force, a first electromagnetic force generating source that generates a magnetic attraction force that displaces a valve in a valve opening direction according to a control current and a magnetic force that displaces a valve in a valve closing direction. A second electromagnetic force generating source for generating an attraction force in accordance with the control current, a calculating means for calculating an operation speed of the valve before the valve is seated, a first electromagnetic force generating source and a second electromagnetic force generating Control means for controlling a control current supplied to the source. Then, the control means adjusts the switching timing of the control current supplied to the electromagnetic force generation source on the valve suction side so that the operation speed calculated by the calculation means becomes slow.

Here, in the second invention, the control means is configured to supply, at the previous control cycle, a timing at which the control current starts to be supplied to the electromagnetic force generating source on the side for sucking the valve or a timing at which the control current is completed. When the operation speed calculated in the current control cycle is reduced due to the advance of the control cycle, it is preferable to change the timing so as to be earlier than the current set timing. In this case, in the previous control cycle, the timing at which the control current is started to be supplied or the timing at which the control current is finished to be supplied to the electromagnetic force generation source on the side that sucks the valve is advanced, so that it is calculated in the current control cycle If the operating speed increases, the timing is changed so as to be later than the currently set timing.

Further, it is desirable that the control means performs control for displacing the valve in the valve opening direction or the valve closing direction by performing primary overexcitation and secondary overexcitation following primary overexcitation. In this case, in the previous control cycle, the timing at which the control current of the secondary overexcitation is started to be supplied to the electromagnetic force generation source on the side that suctions the valve is advanced,
When the operation speed calculated in the current control cycle becomes slow, the timing may be changed so as to be earlier than the current set timing. Then, in the previous control cycle, the timing at which the control current for secondary overexcitation is started to be supplied to the electromagnetic force generating source on the valve suction side is advanced, so that the operating speed calculated in the current control cycle is obtained. Is changed, the timing is changed so as to be later than the currently set timing.

On the other hand, in the above-described configuration, the control means adjusts the timing of starting to supply the control current before starting to adjust the timing of terminating the supply of the control current to the electromagnetic force generating source on the suction side of the valve. It is preferable to carry out.

It is also preferable that the control means adjusts the timing of the control current in the primary overexcitation, and then adjusts the timing of starting to supply the control current of the secondary overexcitation.

In the above configuration, it is preferable to further provide a lift sensor for detecting a lift position of the valve. In this case, the calculating means calculates the operating speed based on the time required for the valve to pass between a plurality of lift positions near the seating position of the valve.

Further, the calculating means corrects the output of the lift sensor based on the output of the lift sensor when the valve is fully seated and the output of the lift sensor when the valve is fully seated. It is desirable.

[0012]

DESCRIPTION OF THE PREFERRED EMBODIMENTS First, an example of a valve train to which the present invention can be applied and an example of a control system thereof will be schematically described with reference to FIGS. FIG. 1 is a cross-sectional view showing an example of a valve operating device, in which a valve 1 is interposed at an intake port and / or an exhaust port of each cylinder of an engine. This valve train has a pair of electromagnetic force generation sources arranged opposite to each other, and the electromagnetic force generated by each of the generation sources causes the lift amount of the valve 1 (intake valve or exhaust valve) to be reduced. Change.

The stem of the valve 1 is slidably inserted into a valve stem guide 3 mounted on a cylinder head 2. Also, on the upper part of the cylinder head 2, 2
A substantially cylindrical assembly composed of two stator cores 3 and 4 and an armature 5 and the like is mounted, and the central axis of this assembly coincides with the central axis of the valve stem guide 3. Stator core 4 (or 5), which is a source of electromagnetic force
Is a magnetic body having a cylindrical shape, and an electromagnetic coil 7 (or 8) is housed in an annular groove formed on one surface thereof. A valve-opening stator core 4 that generates an electromagnetic force that attracts the valve 1 in the valve-opening direction (downward in the figure) is attached to the cylinder head 2. On the other hand, the valve-closing stator core 5 that generates magnetic attraction in the valve-closing direction (upward in the drawing) is integrated with the valve-opening stator core 4 via a lift adjuster 9. The distance between the two stator cores 4 and 5 (that is, the distance at which the valve 1 can be displaced) can be adjusted by the adjuster 9. An armature 6 is arranged in a space between the two stator cores 4 and 5. This armature 6
It is a magnetic material that integrates a disk and a stem. Since the stem of the armature 6 is inserted into the guide hole formed in the case 10 of the valve gear, the armature 6
It can move only in the vertical direction in FIG.

Above and below the armature 6, a valve closing spring 11 and a valve opening spring 12 are arranged, respectively. The valve-closing spring 11 is housed in a space formed by an annular cutting portion (receiving seat) formed on the upper plane of the cylinder head 2 and a through hole of the stator core 4. The spring 11 constantly urges the valve 1 in the closing direction by pushing up a retainer 13 attached to the stem of the valve 1. On the other hand, the valve opening spring 12 is housed in a space formed by a receiving seat formed in the case 10 and a through hole of the stator core 5. This spring 12 pushes the armature 6 downward,
The valve 1 is constantly biased in the opening direction. Armature 6
When no electromagnetic force is applied to the armature 6, the armature 6 is stationary at a neutral position where the urging force of the valve closing spring 11 and the urging force of the valve opening spring 12 are balanced.
In this neutral state, the armature 6 presses the valve 1 downward, and the valve 1 is stationary in a half-open state (half lift). A shim 14 for adjusting the clearance between the armature 6 and the valve 1 is attached to the end of the stem of the valve 1.

The case 10 is provided with a lift sensor 15 for detecting the lift position (lift amount L) of the valve 1. As the lift sensor 15, for example, the needle 1 attached to the tip of the stem of the armature 6 is used.
6, a gap sensor that detects an eddy current that changes according to the displacement can be used. By amplifying the non-linear minute voltage output from the lift sensor 15 by the amplifier 17 and then linearizing by the linearizer 18,
The lift sensor voltage V can be obtained. The lift sensor voltage V is used when adjusting the switching timing of primary overexcitation and secondary overexcitation, which will be described later, and is also used for calculating the seating speed of the valve 1. If necessary, the cylinder head 2 in the vicinity of the valve gear may be used.
A vibration meter 19 may be provided on the upper part of the table (the significance of providing the vibration meter 19 will be described later).

FIG. 2 shows the lift amount L and the lift sensor voltage V
It is a figure for explaining the relation with. Here, the “lift amount” is defined as a displacement amount of the valve 1 with respect to a fully closed position (a state of sitting on a valve seat), and is 0 at a fully closed position and a maximum Lmax at a fully opened position (full lift).
Further, the relationship between the lift amount L and the lift sensor voltage V can be basically expressed as a linear function expression (L = −aV + b). Accordingly, the lift sensor voltage V becomes maximum when the valve 1 is fully closed, decreases linearly with an increase in the lift amount, and becomes minimum when the valve 1 is fully opened.

The basis of the opening and closing movement of the valve 1 is that the current supplied to the pair of electromagnetic coils 7 and 8 is controlled so that the generation / extinction of the electromagnetic force in both is alternately switched.
However, in actual control, it is necessary to perform more complicated control due to problems related to collision noise between members when seated and eddy current. If the velocity of the valve 1 when it is seated (when it is fully opened and when it is fully closed) is large, the collision energy between the armature 6 and the stator cores 4 and 5 becomes large, so that problems such as noise (hitting sound) and damage to members are caused. Occurs. Further, due to the influence of the eddy current generated at the time of high-speed operation of the valve 1, problems such as a decrease in responsiveness of the valve 1 and an increase in power consumption occur. In view of such a problem, in a process of sucking the valve 1 in one direction (for example, a process of displacing the valve 1 from a fully open position to a fully closed position), a method of executing a plurality of excitation modes is effective.

FIG. 3 is a schematic block diagram of a valve control system. Although FIG. 1 shows only a control system related to one valve operating device, in actuality, the same configuration exists in the subsequent stage of the valve control unit 30 by the number of valves.
The valve controller 30 outputs various timing signals to the valve closing valve drive circuit 31 and the valve opening valve drive circuit 32 based on an input signal including the lift sensor voltage V. Timing signals given to the valve closing valve drive circuit 31 include a primary overexcitation timing signal FTS ₁ , a secondary overexcitation timing signal STS ₁ , a reverse excitation timing signal RTS _1, and a hold timing signal HTS ₁ . Similarly, the timing signals given to the valve opening valve drive circuit 32 include a primary over-excitation timing signal FTS ₂ , a secondary over-excitation timing signal ST
S ₂ , a reverse excitation timing signal RTS ₂ and a hold timing signal HTS ₂ . Each timing signal FTS, STS,
RTS and HTS indicate the timing of each excitation (start timing / end timing regarding supply of control current). Further, the reverse excitation timing signal RTS includes an instruction such that the energization direction of the electromagnetic coils 7 and 8 is reversed. Further, the hold timing signal HTS is a PWM signal having a predetermined frequency (for example, 1 kHz), and its duty ratio corresponds to the hold current I _hold . When the PWM signal flows through the electromagnetic coils 7 and 8, the waveform becomes dull due to the inductance of the electromagnetic coils and becomes a substantially constant current value.

FIG. 12 is a control block diagram of the valve control section 30. The timing generator 52 generates timing signals FTS, STS, RT for instructing the switching timing of each excitation based on the output signal of each comparator 51.
Generate S and HTS. The state of the timing signal that defines the rise / fall of the primary overexcitation and the rise of the secondary overexcitation changes at the timing when the output of the comparator 51 changes. A lift sensor voltage V indicating the lift amount of the valve 1 is input to the non-inverting input terminals of all the comparators 51. On the other hand, the inverting input terminal of each comparator 51 is connected to each output channel of the DA converter 53, and different trigger levels (set voltages) are input. Here, the “trigger level” is a threshold voltage related to the lift sensor voltage V, and is a value for detecting that the valve 1 has reached a specific lift position. As described above, since the lift amount L of the valve 1 and the lift sensor voltage V are in a linear relationship, designating the trigger level means designating the lift amount L of the valve 1. In one comparator 51 to which the predetermined trigger level and the lift sensor voltage V are input, when the output state changes from the L level to the H level, the valve 1 reaches the position specified by the trigger level. It corresponds to the time when it was done. Each of the trigger levels defining the excitation timing is calculated by the microcomputer 50 (details will be described later). The trigger level data calculated by the microcomputer 50 is transferred to the DA converter 53 together with the channel data. The DA converter 53 outputs a trigger level which is an analog signal corresponding to the trigger level data from an output channel specified by the channel data.

FIG. 4 is a schematic circuit diagram of the valve drive circuit 31 for closing the valve. Note that the configuration of the valve opening valve drive circuit 32 is the same except that the input timing signal is different. FIG. 5 is a timing chart of valve control. The booster circuit 41 generates a high voltage (for example, 100 V) by boosting a power supply voltage (for example, 12 V). The generated high voltage is applied to the capacitors C1, C2, and C3 via the diodes, and electric charges are accumulated in each capacitor according to the capacitance. As an example, capacitors C1, C2, C3
Have capacitances of 100 μF, 20 μF, and 10 μF, respectively. Timing signals FTS ₁ and STS from the valve control unit 30
₁ , HTS ₁ , RTS ₁ , _one of the transistors T 1, T 2, T 3 is turned on, and a control current flows through the valve closing electromagnetic coil 8.

When closing the valve 1 in the open state,
First, during a period from time t1 when the lift sensor voltage V matches the trigger level V11 to time t2 when it matches the trigger level V12, primary overexcitation is performed as the first stage of the valve closing process. Here, “primary overexcitation” refers to an excitation mode in which the valve 1 is accelerated with a large magnetic attraction force at the beginning of rising. During this excitation period, the primary over-excitation timing signal FTS ₁ becomes H level, and the reverse excitation timing signal RTS ₁
Remain at the L level. Therefore, the transistor T1 is turned on, the capacitor C1 is discharged, and a current flows from the node A1 connected to the valve opening electromagnetic coil 8 to the node B1 (the transistor T5 is turned on and the transistor T6 is turned off). Since the capacity of the capacitor C1 is much larger than the capacitors C2 and C3, the current flowing through the electromagnetic coil 8 increases rapidly. At this time, the magnetic attraction of the valve opening electromagnetic coil 7 has almost disappeared, and the valve 1 is accelerated toward the valve closing side by the magnetic attraction in the valve closing direction. When the valve 1 reaches a predetermined lift amount (the lift sensor potential V becomes V2
), The transistor T1 is turned off, so that the magnetic attraction decreases after time t2.

Next, during the period from time t3 when the lift sensor voltage V matches the trigger level V13 to time t4 when the lift sensor voltage V matches the trigger level V14, secondary overexcitation is performed after primary overexcitation. Here, the “secondary overexcitation” refers to an excitation mode for adjusting the seating speed, which is performed after the primary overexcitation.
During this excitation period, the secondary over-excitation timing signal STS ₁ becomes H
Level, and the reverse excitation timing signal RTS ₁ remains at L level. Therefore, the transistor T2 is turned on, the capacitor C2 is discharged, and a current flows from the node A1 connected to the valve closing electromagnetic coil 8 to the node B1. Capacitor
Because the capacitance of C2 is very small compared to capacitor C1,
The magnetic attraction generated by the electromagnetic coil 8 is not as great as the primary overexcitation.

During the period from time t5 when the lift sensor voltage V matches the trigger level V15 immediately before the seating position of the valve 1 to a synchronization point t6 with the crank angle, the hold current I
By flowing _hold , the valve 1 stops in a fully closed state. During this period, the hold timing signal HTS ₁ becomes H level, and the reverse excitation timing signal RTS ₁ remains at L level. Therefore, the transistor T4 turns on and the node A1
, A current flows in the direction of the node B1. After a series of valve closing processes including a plurality of excitation modes, the valve 1 is almost fully closed, so that the current value required to hold the valve 1 may be used. Hold timing signal HTS ₁ as described above is a PWM signal, the pulse width is variable. By supplying such a signal to the base of the transistor T3, the coil current has a value corresponding to the duty ratio of the hold timing signal HTS _1.

After time t6, the valve-opening side is excited to open the valve 1 in the fully closed state. Similarly to the above-described valve closing control, primary overexcitation is performed in a period from time t7 to time t8,
Perform secondary overexcitation period of time t10 from the time t9, the and starts supplying the hold current I _{h old} from the time t11. At this time, reverse excitation on the valve closing side is performed for a predetermined time (time set by the timer) from time t6 when supply of the hold current I _hold is stopped. Here, "reverse excitation" means that when the valve 1 held in the seated position is released, the direction of energization of the electromagnetic coil 8 on the release side is momentarily reversed to reduce the generated eddy current. Refers to the excitation mode to be set. Specifically, in this period, the transistor T3 is then turned on by the inverse excitation timing signal RTS ₁ to H level, the capacitor
Discharge C3. At this time, since the transistor T5 is off and the transistor T6 is on, the current of the valve closing electromagnetic coil 8 flows from the node B1 to the node A1. That is,
Since a reverse current according to the capacitance of the capacitor C3 (the capacitance of which is smaller than C1 and C2) flows through the electromagnetic coil 8, the magnetic flux change due to the current change is canceled out, and the magnetic attraction force remaining on the valve closing side disappears. Let me.

Next, an outline of the control of the seating speed according to the present embodiment will be described with reference to FIG. 6 and FIG. FIG. 6 is a diagram showing the relationship between the time T and the lift amount L when the secondary overexcitation is set to the same state and only the intensity of the magnetic attraction force of the primary overexcitation is changed. When an appropriate magnitude of magnetic attraction is generated in the primary overexcitation, the lift amount L of the valve 1 changes as shown by a solid line a due to the magnetic action. That is, the speed of the valve 1 (Δ
L / ΔT) is sufficiently decelerated near the sitting position. In this state, since the suction by the secondary overexcitation has started, the seating speed is sufficiently low (the seating speed is set to υa
), And there is almost no hitting sound during sitting. On the other hand, when the magnetic attraction force due to the primary overexcitation is too strong, the displacement of the valve 1 changes as shown by the solid line b. That is, the valve 1 is excessively accelerated by the primary overexcitation, and the seating speed υb (> υa) increases. Therefore, a loud tapping sound is generated by the collision energy according to the sitting speed. Conversely, when the magnetic attraction force due to the primary overexcitation is too weak, the displacement of the valve 1 changes as shown by the solid line c. In this case, since the magnetic attraction force due to the primary overexcitation is too weak, after the speed becomes 0 at a position away from the seating position, the biasing force of the valve-opening spring 12 causes the direction opposite to the seating position (this direction). In this case, it starts to move to the fully open side. Since the secondary overexcitation is started while the valve 1 is accelerating to the opposite side, the valve 1 is excessively accelerated by the magnetic attraction of the secondary overexcitation, and the magnetic attraction of the primary overexcitation is performed. Is too strong, the seating speed υc (>
υa) becomes faster. As a result, also in this case, a loud tapping sound is generated at the time of sitting. It can be seen from the above case that the seating speed increases when the magnetic attraction force due to primary overexcitation is not appropriate, that is, whether the magnetic attraction force is too large or too small. .

If the magnitude of the magnetic attraction force of the primary overexcitation is adjusted by feedback control based on the seating speed based on such knowledge, the seating speed can be converged within an appropriate range. That is, when the seating speed is increased due to the magnetic attraction force of the primary overexcitation being too weak (the case of the solid line c), an adjustment is made to increase the magnetic attraction force of the primary overexcitation. Just do it. Also,
In the case where the seating speed is increased due to the magnetic attraction of the primary overexcitation being too strong (the case of the solid line b), an adjustment may be made to weaken the magnetic attraction of the primary overexcitation. Good. However, it is not possible to specify which of the solid lines b and c the valve 1 has traced from the detected seating speed itself. Therefore, if the magnetic attraction force is changed in one direction (for example, on the increasing side), and the seating speed is slowed down (in the case of the solid line c), the magnetic attraction force may be changed in the same direction the next time. Conversely, if the seating speed is increased as a result of increasing the magnetic attraction force (the case indicated by the solid line b), it may be changed in the opposite direction (ie, on the decreasing side) next time.

The magnetic attraction in the primary overexcitation can be changed by the following methods (1) to (3). (1) Adjustment of startup timing of primary overexcitation Startup timing Magnetic attraction of primary overexcitation Advance Increase Increase Delay Decrease (2) Adjustment of fallover timing of primary overexcitation Falling timing Magnetic of primary overexcitation Attraction force Advance Decrease Delay Increase (3) Coil supply current in primary overexcitation Supply current Magnetic attraction force of primary overexcitation Large Increase Small Decrease Here, "magnetic attraction force" is the influence force to displace the valve. It is defined and basically determined by the magnitude of the generated electromagnetic force itself (that is, the current value) and the period of generation. Therefore, even if an electromagnetic force of the same magnitude is generated, the magnetic attraction becomes smaller as the period is shorter.

FIG. 7 shows that the primary overexcitation is set to the same state,
FIG. 9 is a diagram showing a relationship between time T and lift amount L when only the rising timing of secondary overexcitation is changed. When the rising timing of the secondary overexcitation is appropriate, the lift amount L of the valve 1 changes as shown by the solid line a by the magnetic action. That is, in the vicinity of the seating position,
At time ta when the valve 1 is sufficiently decelerated, the suction by the secondary overexcitation is started. Therefore, the seating speed is sufficiently low, and a striking sound during sitting is hardly generated. On the other hand, if the startup timing of the secondary overexcitation is too early (time tb),
The valve 1 changes as shown by the solid line b. That is,
Before the valve 1 is sufficiently decelerated, a magnetic attraction force due to the secondary overexcitation acts, so that the valve 1 is excessively accelerated and the seating speed is increased. Conversely, if the rising timing of the secondary overexcitation is too late (time tc), the valve 1 changes as shown by the solid line c. That is, in a state where the valve 1 is accelerating in the direction opposite to the seating direction by the urging force of the valve-opening spring 12, magnetic attraction by the secondary overexcitation acts. As a result, the valve 1 is excessively accelerated by the magnetic attraction force of the secondary overexcitation, and the seating speed increases. It can be seen from the above case that the seating speed increases when the rising timing of the secondary overexcitation is not appropriate, that is, whether the timing is too early or too late.

By adjusting the start-up timing of the secondary overexcitation by feedback control based on the seating speed based on such knowledge, the seating speed can be kept within an appropriate range. That is, if the seating speed is increased due to too early a start-up timing of the secondary overexcitation (the case of the solid line b), adjustment may be made to delay the timing. Further, when the seating speed is increased due to too late start-up timing of the secondary overexcitation (the case indicated by the solid line c), adjustment may be made to advance the timing. However, it is not possible to specify which of the solid lines b and c the valve 1 has traced from the detected seating speed itself. Therefore, first, when the magnetic attractive force is changed in one direction (for example, a direction in which the magnetic force is accelerated), and the seating speed is thereby reduced (the case indicated by the solid line c), the magnetic force may be changed in the same direction the next time. Conversely, if the seating speed becomes faster as a result of the earlier start-up timing (the case of the solid line b), it may be changed in the opposite direction (that is, in the direction of delay) next time.

Following the conceptual description of the seating speed control according to the present embodiment, more specific control will be described below in detail. FIG. 8 is a flowchart showing a basic routine for controlling the seating speed of the valve 1. The microcomputer 50 in FIG. 12 repeatedly executes this routine at a predetermined control cycle. First, based on the lift sensor voltage V read in step 1, the seating velocity control unit 55 calculates the seating velocity upsilon _n of the valve 1 in the present control cycle (n-th) (Step 2). Seating speedυ
_n is both the seating speed on the valve-closing side (the speed immediately before the armature 6 contacts the valve-closing stator core 5) and the seating speed on the valve-opening side (the speed just before the armature 6 contacts the valve-opening stator core 4). Is calculated. Seating speedυ
_n can be calculated from the time required for the valve 1 to pass between two preset points near the seating position.

FIG. 13 is a diagram for explaining a method of calculating the seating speed on the valve opening side. First, two different lift positions (reference positions) are determined in the vicinity of the seating position on the valve-opening side (that is, the fully open position), and trigger levels corresponding thereto are set as reference trigger levels Vbase21 and Vbase22. The set reference trigger level Vbase is
0 is stored in the ROM. At time T1 the valve 1 has reached the one of the reference position (farther from the seating position), the output of the non-inverting input comparator 54 ₁ the trigger level Vbase21 being input to the terminal is switched from L level to H level. At time T2 the valve 1 reaches the other reference position (closer to the seated position), the output of the non-inverting input comparator 54 ₂ the trigger level Vbase22 being input to the terminal is switched from L level to H level . Therefore, the distance DIST between two reference positions, is divided by the time ΔT between the time T1 and the time T2, it is possible to calculate the seating velocity upsilon _n of the valve-opening side. Since the distance DIST is a constant value, the time ΔT (more precisely, 1 / ΔT) may be used as the seating speed υ. In the same manner, two reference positions are determined in the vicinity of the seating position on the valve-closing side, and the trigger levels corresponding to the two reference positions are set to the reference trigger levels Vbase11 and Vbase12.
Is set as, from the output of the comparator _543, 54 _4, it is possible to calculate the seating velocity upsilon _n of valve-closing side. In addition, the seating speed calculated in the current control cycle υ
_n is, when a sudden change relative to the value upsilon _n-1 calculated in the last control cycle _{(e.g., υ n ≧ υ n-1} ± 10%), this value upsilon
without adopting the _n, it may be adopted a previous value υ _n-1. Alternatively, a moving average value of the past seating speed (for example, an average value of the present value and the past five points) may be adopted.

[0032] Based on the seating velocity upsilon _n of the thus calculated valve opening side (or the valve closing side), the excitation timing controller 56 calculates the excitation timing of the valve opening side (or the valve closing side) Perform (Step 3). As described above, since the excitation timing is specified by the trigger level, the processing in this step corresponds to calculating each trigger level V _{n + 1} in the next control cycle. The trigger level Vn _{+ 1} calculated by such feedback control is output to the DA converter 53 together with the channel data as a new trigger level (step 4).

(Calculation of startup timing of primary overexcitation)
An example of the process of step 3 performed by the excitation timing control unit 56 will be described with reference to FIG. FIG. 9 is a flowchart showing a subroutine for calculating the timing of starting the primary overexcitation. First, in step 11, whether the seating velocity upsilon _n calculated in Step 2 in FIG. 8 is equal to or smaller than the threshold Upushironth, i.e., whether the current seating velocity upsilon _n converges to within the permissible range Is determined. In one control period, if an affirmative determination is made in step 11, it is not necessary to change the current seating velocity upsilon _n. Therefore, the current trigger level V related to the rising timing of the primary overexcitation is maintained even in the next control cycle (step 12). More specifically, regarding the trigger level V11 indicating the timing of starting the primary overexcitation in the valve closing process, the current value V11
_n is set as the next trigger level V11 _{n + 1} .
On the other hand, the current value V21 _n of the trigger level that indicates the rising timing of the primary overexcitation in the valve opening process is set as the next trigger level V21 _{n + 1} . The initial values of these trigger levels V11 and V21 are stored in the ROM constituting the microcomputer 50 (the same applies to trigger levels described later). Step 13 following step 12
In, the initial correction request flag FAMD is set to “1”. Here, the initial correction request flag FAMD is a flag provided when starting the change of the startup timing of the primary overexcitation, so that only the first control cycle goes through an execution procedure different from the subsequent control cycle. .

After a series of steps 11 to 13 described above, this routine in the current control cycle is terminated, and the execution in the next control cycle is awaited. Therefore, as long as the seating speed is appropriately reduced, the series of steps 11 to 13 is executed in each control cycle, so that the start timing of the primary overexcitation is not changed.

Next, in a certain control cycle, if it is determined in step 11 that the seating speed υn is larger than the threshold value 周期_th , the procedure from step 14 is executed to converge the increased seating speed. , The excitation timing is changed. First, a timing correction term a corresponding to the offset amount of the rising timing of the primary overexcitation is calculated (step 14). Timing correction term a (positive number) is accompanied from the seating velocity upsilon _n to a value threshold υth by subtracting Δυ increases, it sets a large value. Accordingly, when the seating speed is high, the seating speed can be quickly converged by applying a large value to the timing correction term a. On the other hand, when the seating speed is low (however, it is larger than the threshold value Δth), the seating speed can be finely adjusted by applying a small value to the timing correction term a. Note that the timing correction term a may be a constant value regardless of the above subtraction value Δυ. Step 14
In step 15 following the above, the initial correction request flag FA
It is determined whether or not MD is “1”. If the start timing of the primary overexcitation has not been changed in the previous control cycle (n-1st time), that is, if the current control cycle (nth time) is the first time change, step Proceed to 20. This is because the flag FAMD has been set to “1” by the processing in step 13 in the previous control cycle. In this case, after the initial correction request flag FAMD is reset to "0" (step 20), a calculation for accelerating the rising timing of the primary overexcitation is performed (step 18).

[0036] That is, with respect to the trigger level V11 to direct the rise timing of the primary overexcitation in closing process, the value obtained by subtracting the timing correction term a calculated from the current value V11 _n at step 14, the next trigger It is set as level V11 _{n + 1} (step 18). As a result, the next trigger level V11 _{n + 1} becomes the current value V11 _n
Therefore, the rising timing (time t1 in FIG. 5) of the primary overexcitation in the valve closing process is advanced. On the other hand, as for the trigger level V21 which indicates the rising timing of the primary overexcitation in the valve opening process, the value obtained by adding the timing correction term a calculated in step 14 to the current value V21 _n is the next trigger level V21.
It is set as _{n + 1} (step 18). As a result, the next trigger level V21 _{n + 1} becomes larger than the current value V21 _n , so that the rising timing of the primary overexcitation in the valve opening process (time t7 in FIG. 5) is advanced. As a result of hastening the rising timing of the primary overexcitation at the time of opening / closing the valve, the magnetic attraction force for attracting the valve 1 in this excitation mode is greater than at present.

The first timing advance change as described above (step 18 after steps 11, 14, 15, 20)
Is appropriate, the seating speed decreases. The case where the seating speed is reduced by advancing the rising timing of the primary overexcitation is the case where the valve 1 traces the solid line c in FIG. 6, that is, the case where the magnetic attraction force due to the primary overexcitation is too weak. . In this case, the rising timing of the primary overexcitation is gradually advanced (the magnetic attraction force is increased) until the seating speed properly converges according to the procedures of steps 11, 14, 15 to 18. .

In the control cycle in which the first timing is changed earlier, the initial correction request flag FAMD is reset to "0" (step 20). Therefore, in each control cycle thereafter, the procedure after step 16 is executed. First, in step 16, the current seating speed υ _n
Is determined to be lower than the immediately preceding sitting speed υ _n-1 . In the case of the solid line c in FIG. 6, the seating speed is reduced this time as compared with the previous time by advancing the rising timing of the primary overexcitation. Therefore, the process proceeds from step 16 to step 17, and it is determined whether or not the rising timing of the primary overexcitation has been advanced in the previous control cycle. More specifically, as shown in Table 1, the determination in step 17 is the trigger level V11 in the previous control cycle (the (n-1) th control cycle).
_{This is} made possible by comparing _n-1 (or V21 _n-1 ) with the trigger level V11 _n-2 (or V21 _n-2 ) in the control cycle two times before (n-2 times).

[0039]

[Table 1] Previous primary excitation control V11 (valve-closed side) V21 (valve-opened side) V11 _n-1 <V11 _n-2 V21 _n-1 > V21 _n-2 rise timing V11 _n-1 > V11 _n-2 V21 _n-1 <V21 _n-2

In the previous control, the first timing advance change is performed, and an affirmative determination is made in step 17, so that the above-described calculation for further accelerating the startup timing of the primary overexcitation is performed (step 18).

The above-described series of procedures are repeatedly executed until the seating speed converges to the threshold value Δth or less (step 11). When the seating speed converges, the process proceeds from step 11 to step 12, where the trigger level V11 (or V21) at the time of convergence, that is, the rising timing of the primary overexcitation at that time is maintained.

On the other hand, if the early timing change as described above is not appropriate, the seating speed will be faster than the previous one. The case where the seating speed is increased by advancing the timing of the primary overexcitation is shown in FIG.
In this case, the valve 1 traces the solid line b, that is, the magnetic attraction force due to the primary overexcitation is too strong. In this case, until the seating speed converges properly,
By repeatedly executing steps 11, 14 to 16, (17) and 19, the rise timing of the primary overexcitation is gradually delayed (the magnetic attraction force is reduced).

In the control cycle in which the first timing advance is changed, the initial correction request flag FAMD is reset to "0" (step 20). Therefore, in each control cycle thereafter, the procedure after step 16 is executed. First, in step 16, the current seating speed υ _n
Is determined to be lower than the immediately preceding sitting speed υ _n-1 . In the case of the solid line b in FIG. 6, unlike the case of the solid line c, the seating speed becomes faster this time than the previous time due to the first timing advance change. Therefore, the process proceeds from the determination in step 16 to step 19, and an operation for delaying the rising timing of the primary overexcitation is performed. More specifically, a value obtained by adding the timing correction term a calculated in step 14 to the current value V11 _n with respect to the trigger level V11 indicating the rising timing of the primary overexcitation in the valve closing process is the next trigger. Level V
11 Set as _{n + 1} . Thereby, the next trigger level V11 _{n + 1} becomes larger than the current value V11 _n ,
The rising timing of the primary overexcitation in the valve closing process (time t1 in FIG. 5) is delayed. On the other hand, as for the trigger level V21 which indicates the rising timing of the primary overexcitation in the valve opening process, the value obtained by subtracting the timing correction term a calculated in step 14 from the current value V21 _n is the next trigger level V21. Set as _{n + 1} . As a result, the next trigger level V21 _{n + 1} becomes smaller than the current value V21 _n , so that the rising timing of the primary overexcitation in the valve opening process (time t7 in FIG. 5) is delayed.

In step 19, as a result of delaying the rising timing of the primary overexcitation at the time of opening / closing the valve, in the case of the solid line b in FIG. 6, the magnetic attraction force due to the primary overexcitation decreases and the seating speed is reduced. Also slows down. Therefore, in each of the subsequent control cycles, after the determination in step 16, step 1
Then, it is determined whether or not the rising timing of the primary overexcitation has been advanced in the previous control cycle. In this case, since a negative determination is made in step 17, the calculation for further delaying the rising timing of the primary overexcitation as described above is performed (step 19).

The above series of procedures are repeatedly executed until the seating speed converges to the threshold value Δth or less (step 11). When the seating speed converges, the process proceeds from step 11 to step 12, where the trigger level V11 (or V21) at the time of convergence, that is, the rising timing of the primary overexcitation at that time is maintained.

As described above, the routine shown in FIG. 9 is executed for each control cycle, and the rising timing of the primary overexcitation is adjusted by the feedback control, so that the seating speed can be converged within an appropriate range. . As a result, it is possible to effectively suppress the generation of a tapping sound during sitting.

(Calculation of fall timing of primary overexcitation)
Another example of the process of step 3 in FIG. 8 will be described with reference to FIG. FIG. 10 is a flowchart showing a subroutine for calculating the fall timing of the primary overexcitation. First, in step 31, the threshold is seated speed upsilon _n calculated in Step 2 of FIG. 8 Upushironti
It is determined whether it is less than or equal to h. In one control period, if an affirmative determination is made in step 31, it is not necessary to change the current seating velocity upsilon _n. Therefore, the current trigger levels V12 and V12 related to the fall timing of the primary overexcitation
22 is maintained continuously in the next control cycle (step 32). Then, in a step 33 following the step 32, the initial correction request flag FAMD is set to “1”.
Is set, and the execution of this routine in the current control cycle ends.

Next, at a certain control period, when the seating velocity upsilon _n is determined to be larger than the threshold υth In step 31, the procedure after step 34 is executed, for converging the elevated seating velocity , The excitation timing is changed. First, a timing correction term b corresponding to the offset amount of the fall timing of the primary overexcitation is calculated (step 34). This timing correction term b (positive number)
It is for the same reason as the timing correction term a described above, with the seating velocity upsilon _n to a value threshold υth by subtracting Δυ increases, sets a large value. Note that the above subtraction value Δ
Regardless of υ, the timing correction term b can be a constant value. The fall timing correction term b is
It is preferable to set a value smaller than the rising timing correction term a of the primary overexcitation set at the same Δυ. The reason is that the adjustment of the fall timing of the primary overexcitation has a greater effect on the seating speed than the adjustment of the rise timing of the primary overexcitation. That is, the valve 1 at the fall time t2 (or t8) of the primary overexcitation is more likely to be applied to the stator core 5 (or 4), which is a source of magnetic attraction, than the rise time t2 (or t8) of the primary overexcitation. Close. Therefore, considering the case where the rise timing and the fall timing are shifted by the same amount, the fall timing causes a larger amount of change in the sitting speed due to the magnetic field. Therefore, considering the influence of both on the seating speed, at least Δυ
Correction term b set when is relatively large
Is desirably set to a value smaller than the timing correction term a.

In step 35 following step 34, it is determined whether or not the initial correction request flag FAMD is "1". If the fall timing of the primary overexcitation has not been changed in the previous control cycle (n-1 time), the initial correction request flag FAMD is reset to “0” (step 40), and then the primary overexcitation flag FAMD is reset. A calculation is performed to advance the fall timing of the excitation (step 3).
8). That is, for the trigger level V12 to direct the falling timing of the primary overexcitation in closing process, the value obtained by subtracting the timing correction term b calculated from the current value V12 _n at step 34, the next trigger level V
It is set as 12 _{n + 1} (step 38). Thus, the next trigger level V12 _{n + 1} to become smaller than this value V12 _n, is advanced (time t2 in FIG. 5) falling timing of the primary overexcitation in closed processes. On the other hand, as for the trigger level V22 indicating the fall timing of the primary overexcitation in the valve opening process, the value obtained by adding the timing correction term b calculated in step 34 to the current value V22 _n is the next trigger level V22 _{n It} is set as ₊₁ (step 38). Thereby, the next trigger level V22 _{n + 1} is to become larger than the current value V22 _n, falling timing of the primary overexcitation in opening process (time t8 in FIG. 5) is advanced. As a result of hastening the fall timing of the primary overexcitation at the time of opening / closing the valve, the magnetic attraction force for attracting the valve 1 in this excitation mode is reduced as compared with the present.

The first timing advance change as described above (step 38 after steps 31, 34, 35 and 40)
Is appropriate, the seating speed decreases. The case where the seating speed is reduced by accelerating the fall timing of the primary overexcitation is the case where the valve 1 traces the solid line b in FIG. 6, that is, the case where the magnetic attraction force due to the primary overexcitation is too strong. . In this case, the fall timing of the primary overexcitation is gradually advanced (the magnetic attraction force is reduced) until the seating speed appropriately converges according to the procedures of steps 31, 34, and 35 to 38. .

In the control cycle in which the timing was changed earlier, the initial correction request flag FAMD has been reset to "0" (step 40). Therefore, in each control cycle thereafter, the procedure after step 36 is executed. First, in step 36, the current seating speed υ _n
Is determined to be lower than the immediately preceding sitting speed υ _n-1 . In the case of the solid line b in FIG. 6, the seating speed is lower this time than the previous time by making the fall timing of the primary overexcitation earlier. Therefore, the process proceeds from step 36 to step 37, and it is determined whether or not the fall timing of the primary overexcitation has been advanced in the previous control cycle. Specifically, as shown in Table 1, the determination in step 37 is that the trigger level V12 in the previous control cycle (the (n-1) th control cycle) is used.
_This can be achieved by comparing _n-1 (or V22 _n-1 ) with the trigger level V12 _n-2 (or V22 _n-2 ) in the control cycle (n-2nd) two times before.

[0052]

[Table 2] Previous primary overexcitation control V12 (valve-closed side) V22 (valve-opened side) V12 _n-1 <V12 _n-2 V22 _n-1 > V22 _n-2 fall timing with earlier fall timing V12 _n-1 > V12 _n-2 V22 _n-1 <V22 _n-2

In the previous control, the first timing advance change is performed, and an affirmative determination is made in step 37, so that the above-described calculation for further shortening the fall timing of the primary overexcitation is performed (step 38). A series of procedures as described above are repeatedly executed until the seating speed converges to the threshold value Δth or less (step 31). When the seating speed converges, the process proceeds from the judgment in step 31 to step 32, where the trigger level V
12 (or V22), that is, the fall timing of the primary overexcitation at that time is maintained.

On the other hand, when the valve 1 traces the solid line c in FIG. 6, the seating speed is higher than the previous time. In this case, the steps of steps 31, 34 to 36, (37), and 39 are repeatedly performed until the seating speed appropriately converges, thereby gradually delaying the fall timing of the primary overexcitation (magnetic Target suction force). In the control cycle in which the first timing is changed earlier, the initial correction request flag FAMD is set to “0”.
(Step 40). Therefore, in each control cycle thereafter, the procedure after step 36 is executed. First, in step 36, the current seating speed υ
_It is determined whether or not _n is lower than the immediately preceding sitting speed υn _-1 . In the case of the solid line c in FIG. 6, unlike the case of the solid line b, the seating speed is higher this time than the previous time due to the first timing advance change. Therefore, the process proceeds directly from the determination in step 36 to step 39,
An operation for delaying the fall timing of the primary overexcitation is performed. Specifically, with respect to the trigger level V12 to direct the falling timing of the primary overexcitation in closed process, the current value V12 _n, the value obtained by adding the calculated timing correction term b in step 34, the next trigger Level V
Set as 12 _{n + 1} . As a result, the next trigger level V12 _{n + 1} becomes larger than the current value V12 _n ,
The fall timing of primary overexcitation in the valve closing process (time t2 in FIG. 5) is delayed. On the other hand, with respect to the trigger level V22 to direct the falling timing of the primary overexcitation in opening process, from the current value V22 _n, the value obtained by subtracting the timing correction term b calculated in step 34, the next trigger level V22 Set as _{n + 1} . Thereby, the next trigger level V22 _{n + 1} is to become smaller than this value V22 _n, (time t8 in FIG. 5) falling timing of the primary overexcitation in opening process is slower.

In step 39, as a result of delaying the fall timing of the primary over-excitation at the time of valve opening / closing, the magnetic attraction force due to the primary over-excitation increases in the case of the solid line c in FIG. Also slows down. Accordingly, in each of the subsequent control cycles, the determination in step 36 is followed by step 3
Then, it is determined whether the fall timing of the primary overexcitation has been advanced in the previous control cycle. In this case, since a negative determination is made in step 37, the calculation for further delaying the fall timing of the primary overexcitation as described above is performed (step 39).

The above series of procedures are repeatedly executed until the seating speed converges to the threshold value Δth or less (step 31). When the seating speed converges, the process proceeds from step 31 to step 32, where the trigger level V12 (or V22) at the time of convergence, that is, the fall timing of the primary overexcitation at that time is maintained.

As described above, the routine shown in FIG. 10 is executed for each control cycle, and the fall timing of the primary overexcitation is adjusted by the feedback control, so that the seating speed can be converged within an appropriate range. . As a result, it is possible to effectively suppress the generation of a tapping sound during sitting.

(Calculation of startup timing of secondary overexcitation)
Another example of the process of step 3 in FIG. 8 will be described with reference to FIG. FIG. 11 is a flowchart showing a subroutine for calculating the startup timing of the secondary overexcitation. First, in step 51, the threshold is seated speed upsilon _n calculated in Step 2 of FIG. 8 Upushironti
It is determined whether it is less than or equal to h. In one control period, if an affirmative determination is made in step 51, it is not necessary to change the current seating velocity upsilon _n. Therefore, the current trigger levels V13 and V13 related to the secondary overexcitation start-up timing
23 is continuously maintained in the next control cycle (step 52). Then, the initial correction request flag FAM
D is set to "1" (step 53), and the execution of this routine in the current control cycle ends.

Next, at a certain control period, the seating velocity upsilon _n in step 51 if it is determined to be greater than the threshold Upushironth, the procedure after step 54 is executed. First, the timing correction term c corresponding to the offset amount of the rising timing of the secondary overexcitation is calculated in the same manner as the timing correction term a (step 54). Step 54
In a step 55 following the above, the initial correction request flag FA
It is determined whether or not MD is “1”. If the start timing of the secondary overexcitation has not been changed in the previous control cycle (n-1 time), the initial correction request flag FAMD is reset to "0" (step 6).
0), a calculation is performed to advance the rising timing of the secondary overexcitation (step 58). That is, for the trigger level V13 for closing process, the value obtained by subtracting the timing correction term c from the current value V13 _n is set as the next trigger level V13 _{n + 1} (step 58). on the other hand,
Regarding the trigger level V23 for valve opening, the current value V23 _n
Then, a value obtained by adding the timing correction term c calculated in step 54 to the next trigger level V23 _{n + 1} is set (step 58). If the first timing advance change as described above is appropriate, the seating speed becomes slow. The case where the seating speed is reduced by accelerating the rising timing of the secondary overexcitation is the case where the valve 1 traces the solid line c in FIG. In this case, steps 51, 54, 55 to 55 are required until the seating speed properly converges.
Step 58 is repeatedly executed to gradually advance the rising timing of the secondary overexcitation (decreasing the magnetic attraction force).

The control cycle in which the first timing is changed earlier
In the period, the initial correction request flag FAMD is reset to “0”.
Because it is set (step 60), the current seating speed
Degree _nIs the previous sitting speedυ_n-1Is lower than
Disconnected (step 56). The solid line c in FIG.
To speed up the startup timing of secondary overexcitation
As a result, the seating speed is lower this time than in the previous time. Obedience
Therefore, the process proceeds from the determination in step 56 to step 57,
In the previous control cycle, set the startup timing of secondary overexcitation
It is determined whether it has been advanced. The judgment of step 57
Physically, as shown in Table 1, the previous control cycle (n-1 times)
Eye) trigger level V13_n-1(Or V23_n-1)
And the trigger level in the control cycle (n−2 times) two times before
Bell V13_n-2(Or V23_n-2) By comparing
It becomes possible.

[0061]

[Table 3] Previous control of secondary over-excitation V13 (valve-closed side) V23 (valve-opened side) V13 _n-1 <V13 _n-2 V23 _n-1 > V23 _n-2 with early start-up timing V13 _n-1 > V13 _n-2 V23 _n-1 <V23 _n-2 with delayed timing

In the previous control, the first timing advance change is performed, and an affirmative determination is made in step 57, so that the above-described calculation for further accelerating the startup timing of the secondary overexcitation is performed (step 58). . Such a series of procedures is repeatedly executed until the seating speed converges to the threshold value Δth or less (step 51). When the seating speed converges, the process proceeds from step 51 to step 52, where the trigger level V13 at the time of convergence is set.
(Or V23), that is, the rising timing of the secondary overexcitation at that time is maintained.

On the other hand, when the valve 1 traces the solid line b in FIG. 7, the steps 51, 54 to 56, (57) and 59 are repeatedly executed until the seating speed converges. The start timing of the next overexcitation is gradually delayed. That is, in the control cycle in which the first timing is changed earlier, the initial correction request flag FAM
D has been reset to "0" (step 60). Therefore, in each control cycle thereafter, the procedure from step 56 is executed. First, in step 56, whether the current seating velocity upsilon _n drops below the seating velocity upsilon _n-1 immediately preceding it is determined. In the case of the solid line b in FIG. 7, unlike the case of the solid line c, the first timing advance change causes the seating speed to be higher this time than in the previous time. Accordingly, the process proceeds directly from the determination of step 56 to step 59, and an operation for delaying the rising timing of the secondary overexcitation is performed. Specifically, with respect to the trigger level V13 for closing process, the current value V13 _n, a value obtained by adding the timing correction term c calculated in step 54 is set as the next trigger level V13 _{n + 1} . On the other hand, regarding the trigger level V23 for the valve opening process, the current value V
A value obtained by subtracting the timing correction term c calculated in step 54 from 23 _n is set as the next trigger level V23 _{n + 1} .

In step 59, as a result of delaying the rising timing of the secondary overexcitation at the time of opening / closing the valve, the seating speed is reduced in the case of the solid line b in FIG. Therefore,
In each control cycle thereafter, the process proceeds to step 57 after the determination in step 56, and it is determined whether or not the rising timing of the secondary overexcitation has been advanced in the previous control cycle.
In this case, since a negative determination is made in step 57,
The calculation for further delaying the rising timing of the secondary overexcitation as described above is performed (step 59). Such a series of procedures is repeatedly executed until the seating speed converges to the threshold value Δth or less (step 51). When the seating speed converges, the process proceeds from step 51 to step 52, where the trigger level V13 (or V23) at the time of convergence is maintained.

As described above, the routine shown in FIG. 11 is executed at each control cycle, and the rising timing of the secondary overexcitation is adjusted by the feedback control, so that the seating speed can be converged within an appropriate range. it can. As a result, it is possible to effectively suppress the generation of a tapping sound during sitting.

When the adjustment of the rising timing of the primary overexcitation, the adjustment of the fall timing of the primary overexcitation, and the adjustment of the startup timing of the secondary overexcitation are performed in combination, the influence on the operating speed of the valve 1 is not affected. Considering these, it is desirable to prioritize these adjustments. Assuming that the magnetic attraction generated at the time of the rise and the fall of the primary over-excitation is substantially the same, the adjustment of the fall timing executed in a state where the valve 1 is close to the seating position is better. The effect on the seating speed increases. Therefore, in controlling the seating speed, first, fine adjustment is performed on the rising timing side, and only when the seating speed cannot be appropriately reduced by the adjustment, adjustment on the falling timing side having a relatively large effect is started. For the same reason, the adjustment of the fall timing of the primary over-excitation is performed before the adjustment of the rise timing of the secondary over-excitation.

(Correction of Reference Trigger Level Vbase) As can be understood from the above description, since the control of the excitation timing in the present embodiment is executed based on the seating speed of the valve 1, the reference trigger serving as a calculation reference is provided. Level Vbase
It is important to give exactly. In view of such a point, in the present embodiment, an appropriate correction considering the influence of the disturbance element is performed on the reference trigger level Vbase. This correction is performed for each control cycle by the correction unit 58 in FIG. The correction unit 58 outputs the reference trigger level Vbase (Vbase11, Vbase12, Vbase) stored in the ROM.
21, the Vbase22), a value obtained by multiplying the correction coefficient Aamd _n calculated in the control cycle, as the next reference trigger level V'base _{n + 1,} and outputs to the DA converter 53. Correction coefficient Aamd _n, based on the following equation, the correction coefficient calculation unit 57
Is calculated. Here, Vcls _n is the lift sensor voltage in the fully closed state in the current control cycle, and Vopn _n is the lift sensor voltage in the fully open state in the current control cycle. LIFT is a full lift constant (positive number), that is, a design voltage difference when the valve 1 is displaced between fully closed and fully open.

Aamd _n = (Vcls _n −Vop _{n n} ) / LIFT

[0068] All closed - the measured value of the lift sensor voltage difference between the fully opened, if an ideal state in which the potential difference LIFT match a design, the correction coefficient Aamd _n is 1, i.e., reference trigger level Vbase is as reference Trigger level V'bas
e _{n + 1} . However, in practice, the two do not always match because of the influence of manufacturing variations of the lift sensor or aging. Therefore, a value obtained by multiplying the correction coefficient Aamd _n calculated based on the above equation the reference trigger level Vbase, used as the next reference trigger level V'base _{n +1.} By performing such correction for each control cycle, an accurate reference valve position can be specified, so that the reliability of the calculated seating speed can be further improved.

Lift sensor voltages Vopn, Vcls when seated
(Hereinafter, these are collectively referred to as seating voltages) can be calculated from the lift sensor voltage V and the hold timing signal HTS. As can be seen from the timing chart of FIG. 5, only when the valve 1 is held on the fully closed side, the valve-closing hold timing signal HTS ₁ is in the rising state (actually, PWM corresponding to the hold current).
State). Therefore, the lift sensor voltage V during the hold period indicated by the signal HTS ₁ is
The seating voltage Vcls when fully closed can be used. On the other hand, the seating of the fully open state voltage Vopn is lift sensor voltage V in a state in which the valve-opening hold timing signal HTS ₂ rises
It is. When detecting the seating voltages Vopn and Vcls,
Further, if the following points are considered, the reliability of the detected value can be further improved.

(1) The lift sensor voltage after a predetermined time has elapsed from the rise of the hold timing signal HTS is adopted as the seating voltages Vopn and Vcls. Considering the actual behavior of the valve, when the hold timing signal HTS rises, the valve does not always stop firmly in contact with the stator core and may be slightly displaced (bound phenomenon ). In such a transient state, the lift sensor voltage changes according to the displacement of the valve, and this voltage value is used as the seating voltage Vopn,
It is not preferable to adopt it as Vcls. Therefore, by anticipating the time (for example, 1 ms from the rise of the hold timing signal HTS) until the valve is securely brought into contact, and adopting a value after the elapse of the time, the detected seating voltages Vopn and Vcls can be reduced. Variation can be reduced.

(2) In the hold period, the lift sensor voltages during a period excluding the ignition timing of the ignition plug are adopted as the seating voltages Vopn and Vcls. As shown in FIG. 14, at the ignition timing of the ignition plug, the lift sensor voltage greatly fluctuates due to the influence of noise even though the valve is seated. Therefore, the ignition timing specified from the ignition timing signal (generated by an ECU (not shown)) is masked (for example, 1 ms before and after the ignition timing), and the seating voltages Vopn and Vc from other hold periods.
ls is calculated. This allows accurate seating voltages Vopn and Vc without being affected by noise due to ignition.
You can ask for ls. In the case of a multi-cylinder engine, it is preferable to mask the ignition timing of other cylinders.

(3) The seating voltages Vopn and Vcls are calculated from a plurality of sample values measured during the hold period.
Due to the disturbance element, the measured value varies with only one sample value, and there is a possibility that there is a problem in the reliability of the seating voltages Vopn and Vcls. Therefore, as shown in FIG. 15, a plurality of sample values (for example, 10
Sample), and the average value of them is used as the seating voltage Vopn,
Adopted as Vcls. At this time, the sample value is set to a preset threshold value (for example, a set seating voltage ± 20%).
If it exceeds, the sample value is not adopted and
It is preferable to apply the previously calculated value. This allows
The reliability of the calculated seating voltages Vopn and Vcls can be further improved, and control can be stabilized. Further, a moving average of the seating voltage calculated this time and a plurality of past values (for example, the past five times) may be obtained, and the value may be adopted as the present seating voltage.

(4) When obtaining a plurality of sample values, the sample values are obtained at a sampling period asynchronous with the PWM period of the hold timing signal HTS. As shown in FIG. 16, when a plurality of sample values are sampled at a sampling period synchronized with the PWM period of the hold timing signal HTS, the noise is affected by the PWM control, and each lift sensor voltage V has the same level of noise. May ride almost evenly. In this case, the seating voltages Vopn and Vcls calculated by averaging these sample values are offset from the original values due to the influence of noise. This may cause the same problem even when the sampling period is set to a period that is an integral multiple or a fractional multiple of the PWM period. Therefore, as shown in FIG. 17, it is preferable to obtain a sample value at a period that does not coincide with the PWM period or is not an integral multiple or a fractional multiple thereof, that is, a sampling period that is asynchronous with the PWM period. This allows
The seating voltages Vopn and Vcls can be calculated without being greatly affected by noise due to the PWM control.

(Modifications) The present invention is not limited to the configuration of the above-described embodiment, but can be realized by, for example, the following method. Even when these methods are used, substantially the same operation and effects as those of the above embodiment can be obtained.

(1) Timing Setting Using Time Control In the above embodiment, each excitation timing is set by a trigger level. However, the present invention can also be applied to control in which the excitation timing is set by time control using a counter or the like. For example, when the fall timing of the primary over-excitation is set, a predetermined time T elapses from the time when the output signal of the comparator to which the trigger level specifying the rise timing of the primary over-excitation changes from the L level to the H level is input. The time at which the primary overexcitation was performed is defined as the fall time of the primary overexcitation. Then, depending on the seating speed, the current value of the timing correction term obtained by adding or subtracting a predetermined time T _n, if the next predetermined time T _{n + 1,} it is possible to adjust the seating velocity. By adjusting the predetermined time T through such feedback control, the seating speed can be appropriately controlled. For a specific control routine, basically, the output variables in the flowcharts of FIGS. 8 to 11 may be changed.

(2) Estimation of Seating Speed Using Output of Vibration Meter Calculating the seating speed, which is an input variable of control in the above embodiment, is nothing less than estimating the collision energy at the time of sitting. . From such a viewpoint, it is also possible to estimate the seating speed (in other words, the collision energy) from the output of the vibrometer 19 shown in FIG. 1, and control the seating speed based on the estimated seating speed. In this case, when the output of the vibrometer becomes a predetermined value or more, the timing of each excitation is adjusted. Basically, the control routine in this case can be obtained by changing the input variables in the flowcharts of FIGS.

[0077]

As described above, according to the present invention, the operation speed of the valve before seating is calculated, and feedback control based on the calculated value is performed to adjust the excitation timing and the like such that the seating speed is reduced. It is carried out. Through such control, the seating speed of the valve can be reduced, and the sound of collision between members during seating can be effectively reduced.

[Brief description of the drawings]

FIG. 1 is a sectional view showing an example of a valve train having an electromagnetically driven valve.

FIG. 2 is a relationship diagram between a lift amount and a lift sensor voltage.

FIG. 3 is a schematic block diagram of a valve control system.

FIG. 4 is a schematic circuit diagram of a valve drive circuit.

FIG. 5 is a timing chart of valve control.

FIG. 6 is a diagram showing a relationship between time and lift amount due to primary overexcitation;

FIG. 7 is a diagram showing a relationship between time and lift amount due to secondary overexcitation;

FIG. 8 is a flowchart of a valve seating speed control routine.

FIG. 9 is a flowchart showing a subroutine for calculating a startup timing of primary overexcitation;

FIG. 10 is a flowchart showing a subroutine for calculating a fall timing of primary overexcitation;

FIG. 11 is a flowchart showing a subroutine for calculating a startup timing of secondary overexcitation;

FIG. 12 is a control block diagram of excitation timing.

FIG. 13 is a diagram for explaining a method of calculating a sitting speed.

FIG. 14 is a diagram for explaining the effect of ignition noise on the lift sensor voltage.

FIG. 15 is a diagram illustrating a seating voltage method using a plurality of sample values.

FIG. 16 is a diagram showing a plurality of sample values extracted in a sampling cycle synchronized with a PWM cycle;

FIG. 17 is a diagram showing a plurality of sample values extracted at a sampling cycle asynchronous with a PWM cycle.

[Explanation of symbols]

1 valve, 2 cylinder head,
3 Valve stem guide, 4 Valve opening stator core, 5 Valve closing stator core, 6 Armature, 7 Valve opening electromagnetic coil, 8 Valve closing electromagnetic coil, 9 Lift adjuster, 10 Case, 1
1 valve closing spring, 12 valve opening spring, 13 retainer, 14 shim, 15
Lift sensor, 16 needle, 17 amplifier, 18 linearizer, 19 vibrometer, 30 valve control section, 31 valve drive circuit for closing valve, 32 valve drive circuit for opening valve,
41 booster circuit, 50 microcomputer, 51
Comparator, 52 Timing Generator, 53 DA Converter, 54 Comparator, 55 Seating Speed Calculation Unit, 56 Excitation Timing Control Unit, 57 Correction Coefficient Calculation Unit, 58
Correction unit

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) F16K 31/06 305 F16K 31/06 305S 310 310A 320 320A 385 385A H01F 7/16 H01F 7/18 B 7 / 18 7/16 RF term (reference) 3G092 AA11 DA07 DF05 DG02 DG09 EA00 EA02 EA03 EA04 EA08 EA17 EA22 EA25 EB04 EB05 EB06 EB08 EC01 EC08 FA06 FA14 FA36 FA48 HA12X HA12Z HA13X HA13Z 3G301 JA07 NB07 JA37 NB20 ND01 ND41 NE06 NE11 NE12 NE23 PE10A PE10Z PG02A PG02Z 3H106 DA07 DA25 DB02 DB12 DB26 DB32 DC02 DC17 DD05 EE19 EE20 FA02 FA08 FB43 GC29 KK17 5E048 AA02 AB01 AD07 BA01

Claims (9)

[Claims] A first spring that urges a valve of the internal combustion engine in a valve opening direction and a second spring that urges the valve in a valve closing direction, wherein the valve is driven by an electromagnetic force. A first electromagnetic force generating source for generating a magnetic attraction force for displacing the valve in a valve opening direction in accordance with a control current; and a magnetic attraction force for displacing the valve in a valve closing direction. A second electromagnetic force generating source that is generated in accordance with the control current; a calculating unit that calculates an operating speed of the valve before the valve is seated; a first electromagnetic force generating source and the second Control means for controlling the control current to be supplied to the electromagnetic force generation source, and the control means, wherein the operating speed calculated by the calculation means is slowed down, and the control means is configured to suction the valve. Supply to electromagnetic force source Valve operating apparatus characterized by adjusting the control current. A first spring for urging a valve of the internal combustion engine in a valve opening direction, and a second spring for urging the valve in a valve closing direction, wherein the valve is driven by an electromagnetic force. A first electromagnetic force generating source for generating a magnetic attraction force for displacing the valve in a valve opening direction in accordance with a control current; and a magnetic attraction force for displacing the valve in a valve closing direction. A second electromagnetic force generating source that is generated in accordance with the control current; a calculating unit that calculates an operating speed of the valve before the valve is seated; a first electromagnetic force generating source and the second Control means for controlling the control current to be supplied to the electromagnetic force generation source, and the control means, wherein the operating speed calculated by the calculation means is slowed down, and the control means is configured to suction the valve. Supply to electromagnetic force source Valve operating apparatus characterized by adjusting the switching timing of the control current. 3. The control means hastened a timing at which the control current is started to be supplied or a timing at which the control current is finished to be supplied to the electromagnetic force generating source on the side for sucking the valve in a previous control cycle. Thus, when the operating speed calculated in the current control cycle is reduced, the timing is changed so as to be earlier than a current set timing. Valve train. 4. The control means hastened a timing at which the control current is started to be supplied or a timing at which the control current is finished to be supplied to the electromagnetic force generating source on the suction side of the valve in a previous control cycle. 4. The method according to claim 2, wherein when the operating speed calculated in the current control cycle increases, the timing is changed so as to be later than a current set timing. Valve train. 5. The control means performs control to displace the valve in a valve opening direction or a valve closing direction by performing a primary overexcitation and a secondary overexcitation following the primary overexcitation. In the control cycle, the timing of starting to supply the control current of the secondary over-excitation to the electromagnetic force source on the side that sucks the valve is advanced, so that the operation calculated in the current control cycle is performed. When the speed is reduced, the timing is changed so as to be earlier than the current set timing, and in the previous control cycle, the electromagnetic force generating source on the side that suctions the valve is If the operation speed calculated in the current control cycle is increased by advancing the timing to start supplying the control current of the secondary overexcitation, the timing is reduced. Has been a valve gear according to claim 2, characterized in that changing to be lower than the current setting timing. 6. The control means adjusts the timing of starting to supply the control current before starting to adjust the timing of ending the supply of the control current to the electromagnetic force source on the side that sucks the valve. The valve train according to claim 3, wherein the valve train is executed. 7. The control device according to claim 1, wherein after performing the timing adjustment of the control current in the primary over-excitation, the control means performs timing adjustment to start supplying the control current of the secondary over-excitation. 5. The valve train according to 5. 8. A lift sensor for detecting a lift position of the valve, wherein the calculating means detects a time required for the valve to pass between a plurality of lift positions near a seating position of the valve. The valve operating device according to claim 1, wherein the operating speed is calculated based on the operating speed. 9. The lifting means based on an output of the lift sensor when the valve is fully seated and an output of the lift sensor when the valve is fully seated. 9. The valve train according to claim 8, wherein the output of the sensor is corrected.