JPH0965685A - Current carrying controller of electric motor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To suppress noises and improve acceleration characteristics by a method wherein a first state in which a target acceleration is larger than an actual acceleration and a second state in which a target acceleration is smaller than an actual acceleration are discriminated from each other and the feedback compensation value of the second state is reduced in comparison with that of the first state. SOLUTION: A vibration compensation value and a speed compensation value are added by an adder 16 to a reference current value S4 given by a current waveform generator 15 and the result is converted into an analog voltage Vr2 which is applied to one of the input terminals of an analog comparator 7a. The voltage Vs6 of a signal S6 corresponding to the detection current of a current sensor 2 is applied to the other input terminal of the analog comparator 7a. The comparison result is outputted as a binary signal S71. If the input signal S71 shows Vs6>Vs2, the state is switched to the 'ON-impossible' state and, if Vs6<Vs2, the state is switched to the 'ON- possible' state and, by a timing controller 17c, an independent control system which performs acceleration feedback control is provided. As a result, the feedback compensation value of the second state can be reduced in comparison with that of the first state, so that the acceleration characteristics can be improved and, further, the vibration of the rotor is reduced and the noises can be suppressed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a power supply control for an electric motor, and more particularly to a power supply control device suitable for use in driving a switched reluctance motor.

[0002]

2. Description of the Related Art Switched reluctance motors (hereinafter referred to as
The SR motor) has a very simple structure, for example, as shown in FIG. SR motor rotor R
Is an iron core formed by simply stacking iron plates, and a plurality of teeth (Ra, Rb, Rc, Rd) formed on the outer peripheral surface so as to project toward the stator S surrounding the outer periphery.
have. The stator has a plurality of teeth (Sa, Sa, formed on the inner peripheral surface of an annular core formed by stacking iron plates).
Sb, Sc, ...) Concentrated winding of electric coils. In this type of SR motor, by energizing each electric coil, an electromagnet is formed in each tooth portion of the stator, and the rotor is rotated by attracting each tooth of the rotor by the magnetic force of the stator. . Therefore, the rotor can be rotated in a desired direction by sequentially switching the energization state of the coil wound around each tooth of the stator according to the rotational position of each tooth of the rotor.

The prior art relating to this type of SR motor is as follows.
For example, it is disclosed in Japanese Patent Laid-Open No. 1-298940.

[0004]

The SR motor has the advantages that it has a simple structure, is mechanically tough, and can operate at high temperatures, but it is practically practical. The reality is that it is not used. One of the causes is that the noise generated during rotation is large.

In the SR motor, when each pole of the rotor is at a specific rotation position, the energization of each pole of the stator is switched on / off. Therefore, at the time of switching, the magnetic attraction force applied to the rotor is changed. The size changes rapidly. Therefore, relatively large mechanical vibrations occur in the rotor and the stator. This vibration produces noise.

In order to suppress such vibrations, the motor
It is necessary to drive the motor as smoothly as possible by controlling the energization of the motor. It is impossible to drive the motor smoothly when only the simple pulsed energization that is generally used is turned on / off. Therefore, it is necessary to finely adjust the current value for each rotation angle of the motor.

One of the main causes of the noise of the SR motor is the fluctuation of the actual acceleration in the rotation direction of the rotor. Therefore, if the actual acceleration in the rotation direction of the rotor is detected and feedback control is performed so as to suppress the change, and the bias amount, that is, the current flowing in the coil is controlled, SR It is possible to reduce the noise of the motor. However, performing such feedback control means suppressing the acceleration of the rotor. Therefore, performing feedback control causes a disadvantage other than noise. For example,
When using an SR motor as a drive source for an automobile,
The drive speed of the SR motor needs to be changed frequently, and the acceleration must be performed quickly. But,
When feedback control is performed to suppress noise, speed change is suppressed, so the responsiveness of the control system to acceleration deteriorates, leading to deterioration in the acceleration characteristics of the vehicle.

As described above, in order to drive the SR motor smoothly, it is necessary to finely adjust the current value for each rotation angle of the motor.

For example, if the following control is performed, the current value can be finely adjusted according to the rotation angle of the motor. That is, a large number of current values (current waveforms) associated with respective rotation angles are stored in a memory in advance, and each time the motor makes a minute rotation, the current value associated with the angle at that time is read from the memory. Then, the control target value of the current is sequentially changed. However, in this type of control, the target value of the current does not depend only on the rotation angle of the motor, but
It must be changed according to the rotation speed of the motor and the required drive torque. A huge memory capacity is required to hold all the current values corresponding to all the changes in many rotation angles, many rotation speeds, and many driving torques of the motor in the memory. Further, it is extremely difficult to store in the memory a current value that can cope with a momentary torque change and a speed change.

Therefore, in order to reduce the capacity of the memory required to hold the current value, the data held in the memory is set to one set corresponding to one kind of rotation speed and one kind of driving torque. It is conceivable that the current value at each rotational position is limited and the calculation processing is performed every time the motor rotational speed or the driving torque changes to rewrite the contents of the memory. However, when such control is performed, it takes time to perform calculation processing and update the contents of the memory, so that the response of the control system becomes very slow. For example,
When a current value corresponding to an angle of 128 steps per waveform of one phase of the coil is required, every time the motor rotation speed or drive torque changes, the number of phases x 128 data is Since it is necessary to perform calculation processing and update of memory contents, it takes a very long time.

It is necessary to compensate for the acceleration control error.
When performing the feedback control, the control system is required to have high responsiveness. However, if it is attempted to update the current target value, that is, the contents of a large number of memories based on the compensation amount generated by the feedback control, it takes a long time to update the memories, so that high responsiveness cannot be expected.

An object of the present invention is to reduce noise generated by an electric motor such as an SR motor and improve acceleration characteristics.

[0013]

In order to solve the above-mentioned problems, the energization control device for an electric motor according to claim 1 is an electric motor.
A motor (1); a motor energizing means (6, 7, 17, 18, 19, 1A) for controlling the energization of the electric motor; an acceleration deciding means for deciding a target acceleration of the electric motor ( 64);
Acceleration detecting means (1d, 7E, 7F) for detecting the actual acceleration of the electric motor; the target acceleration determined by the acceleration determining means, the actual acceleration detected by the acceleration detecting means, and the feedback buck gain. A feedback compensation amount corresponding to the parameter to be determined is generated, and the feedback compensation amount is input to the motor urging means so as to suppress the urging amount of the electric motor. Negative feedback control means (88, 89, 8A); and the magnitude relationship between the target acceleration and the actual acceleration are checked to find a first state in which the target acceleration is greater than the actual acceleration and the target acceleration is less than the actual acceleration. The second state is identified, the parameters are switched between the first state and the second state, and the feedback in the second state is compared with the feedback compensation amount in the first state. Reduce compensation amount To Fi --back gain switching means (87); comprises.

According to a second aspect of the present invention, the acceleration detecting means is a speed detecting means (1d, 7E) for detecting the rotational speed of the electric motor, and a change in the speed detected by the means per unit time. Speed difference detecting means (7) for detecting the amount
F).

An electric motor energization control device according to a third aspect of the present invention is an electric motor (1); an energization map for holding an urging amount target value for each rotational position of a rotor of the electric motor. Means (4
9); Rotational position detecting means (1d) for detecting the rotational position of the rotor of the electric motor; Energizing amount detecting means (2, 3, 4) for detecting the actual energizing amount of the electric motor. ); Based on the rotational position detected by the rotational position detecting means, the energizing amount target value held in the energization map means is input, and the energizing amount target value and the energizing amount detecting means detect the energizing amount target value. Motor energizing means (6, 7, 17, 18, 19, 1A) for controlling energization of the electric motor according to the amount of energy; acceleration determining means for deciding a target acceleration of the electric motor. (64); Map control means (68, 69, 6A) for updating the energizing amount target value held by the energization map means in response to a change in the rotation speed of the electric motor and a change in the target acceleration. Acceleration detecting means (1 for detecting the actual acceleration of the electric motor)
d, 7E, 7F); the target acceleration determined by the acceleration determining means, and the actual acceleration detected by the acceleration detecting means,
Negative feedback control means (88, 89, 8A) for generating a feedback compensation amount according to the parameter for determining the feedback loop gain; and the feedback compensation amount generated by the negative feedback control means, Compensation amount adding means (16) for adding to the input of the motor urging means through a path independent of the map control means; and the magnitude relationship between the target acceleration and the actual acceleration is checked to find out the actual target acceleration. A first state in which the target acceleration is larger than the acceleration and a second state in which the target acceleration is smaller than the actual acceleration are discriminated, and the parameters are switched between the first state and the second state to change the first state. Compared to the amount of feedback compensation in the second state
A feedback gain switching means (87) for reducing the feedback compensation amount is provided.

Further, in claim 4, the acceleration detecting means is speed detecting means (1d, 7E) for detecting the rotation speed of the electric motor, and a change in the speed detected by the means per unit time. Speed difference detecting means (7) for detecting the amount
F).

The symbols shown in the parentheses are reference numerals of corresponding elements in the embodiments described later, but each constituent element of the present invention is a specific element in the embodiments. It is not limited to only.

In the structure of claim 1, the negative feedback control means (88, 89, 8A) includes the target acceleration determined by the acceleration determining means, the actual acceleration detected by the acceleration detecting means, and the feedback loop gain. The feedback compensation amount is generated according to the parameter that determines the
The feedback compensation amount is input to the motor biasing means so as to suppress the biasing amount of the motor. By this feedback control, the biasing amount of the electric motor is controlled so that the deviation between the target acceleration and the actual acceleration is suppressed. That is, assuming a case where the target acceleration is constant, fluctuations in the actual acceleration are suppressed, so that mechanical vibration of the rotor, that is, noise is reduced.

When the feedback compensation amount (negative feedback amount) is increased, the noise reduction effect is further improved, but the response characteristic of the control system, that is, the acceleration performance of the motor is deteriorated. A feedback gain switching means (87) is provided to reduce noise without degrading acceleration performance.

Feedback gain switching means (87)
Examines the magnitude relationship between the target acceleration and the actual acceleration to identify a first state where the target acceleration is larger than the actual acceleration and a second state where the target acceleration is smaller than the actual acceleration.
The parameter is switched between the second state and the second state and compared with the feedback compensation amount in the first state.
In this state, the feedback compensation amount is controlled to be reduced.

For example, in the case of increasing (accelerating) the motor drive speed while keeping the target acceleration constant, the actual acceleration fluctuates in a short cycle as shown in FIG. The state of "" and the state of "target acceleration <actual acceleration" appear alternately. In the former case, the biasing amount (that is, current value) of the electric motor increases in accordance with the feedback compensation amount so that the actual acceleration increases, and in the latter case,
The urging amount of the electric motor is reduced according to the feedback compensation amount so that the actual acceleration is reduced.

When the feedback compensation amount is large, the effect of suppressing the acceleration / deceleration is increased, and the vibration in the rotational direction of the motor is further suppressed, so that the noise is reduced. However, in that case, the time required for speed change is lengthened due to the suppression of acceleration / deceleration. That is, the responsiveness is deteriorated with respect to the change of the acceleration / deceleration target value. Therefore, for example, it is unsatisfactory as a drive system used in an electric vehicle that requires frequent changes in traveling speed according to the driver's will. On the contrary, if the feedback compensation amount is reduced, the effect of suppressing the acceleration / deceleration is reduced, and the responsiveness to the change of the target value of the acceleration / deceleration is improved.
Noise increases.

The feedback compensation amount is determined by the target acceleration, the detected actual acceleration, and the loop gain of the feedback control system. In general, the loop gain is constant, and the feedback compensation amount is determined according to the deviation between the target acceleration and the detected actual acceleration. However, the feedback gain switching means (87) of the present invention.
Is a first state in which the target acceleration is larger than the actual acceleration by examining the magnitude relationship between the target acceleration and the actual acceleration,
A second state in which the target acceleration is smaller than the actual acceleration is discriminated, and the parameters are switched between the first state and the second state to compare the feedback compensation amount in the first state with the feedback compensation amount. The control is performed so as to reduce the feedback compensation amount in the second state. That is, in the present invention, the loop gain of the feedback control system is automatically changed between the first state and the second state. Moreover, the amount of feedback compensation is relatively large in the first state and relatively small in the second state.

In the negative feedback control of the present invention, as shown in FIG. 22, the feedback compensation amount (feedback amount) is large in the first state, and accordingly, the biasing amount of the electric motor ( That is, the current value) increases and the deviation of the actual acceleration from the target value is reduced. Further, in the second state, the feedback compensation amount becomes small, so that the suppression amount of the biasing amount (that is, the current value) of the electric motor becomes small and the acceleration characteristic is improved. Moreover, since the feedback compensation amount can be increased in the first state, a sufficient effect of negative feedback control can be obtained, and vibration of the rotor, that is, noise can be suppressed. These effects have been confirmed by experiments by the present inventor.

In the third aspect, the energization map means (49) holds the urging amount target value for each rotational position of the rotor of the electric motor. Furthermore, the motor urging means (6,
7, 17, 18, 19, 1A) inputs the urging amount target value held in the energization map means based on the rotational position of the rotor of the electric motor detected by the rotational position detecting means. The energization of the electric motor is controlled according to the target value of the bias amount and the actual bias amount detected by the bias amount detecting means.
Therefore, a desired amount of current can be passed through the electric coil of the electric motor for each rotational position of the electric motor. Further, the map control means (68, 69, 6A) updates the urging amount target value held by the energization map means in response to the change in the rotation speed of the electric motor and the change in the target acceleration. As a result, the energization waveform of the electric motor can be finely adjusted according to the driving conditions at that time.
It is possible to drive the motor efficiently, and it is possible to reduce the noise generated when the motor is driven. Moreover, since the energization amount target value held by the energization map means is sequentially updated according to the drive condition, it is not necessary to hold all the energization amount target values under various conditions in the energization map means, The memory capacity can be reduced.

Of course, it takes some time to update the energizing amount target value held by the energization map means, and therefore the control of the energizing amount does not delay the response of the motor to changes in the driving conditions. I can't. However, the feedback compensation amount generated by the negative feedback control means is equal to the compensation amount adding means (1
According to 6), since it is added to the input of the motor energizing means through a path independent of the map controlling means, the negative feedback control is performed regardless of the updating of the energizing amount target value held by the energization map means. It can be carried out, and control with good responsiveness is realized.

In the second and fourth aspects, the speed detecting means (1d, 7E) detects the rotation speed of the electric motor, and the speed difference detecting means (7F) is a unit of the detected rotation speed. The actual acceleration is detected based on the amount of change per unit time.

[0028]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT FIG. 1 shows the configuration of an apparatus according to an embodiment. The device shown in FIG. 1 constitutes a main part of a drive unit of an electric vehicle. In this example, one SR motor 1 is provided as a drive source, and this SR motor 1 is controlled by the controller ECU. Controller E
The CU controls the drive of the SR motor 1 based on the information input from the shift lever, the brake switch, the accelerator switch, and the accelerator opening sensor. Power is supplied from the battery.

The basic structure of the SR motor 1 and its driving principle are shown in FIG. The SR motor 1 shown in FIG. 18 is composed of a stator S and a rotor R rotatably supported in its inner space. The rotor R is formed by laminating a large number of thin iron plates, and has four pole portions Ra, Rb, R projecting outward at positions offset by 90 degrees from each other on the outer circumference.
c and Rd are formed. The stator S is also formed by laminating a large number of thin iron plates, and has six pole portions S protruding inward at positions on the inner circumference that are offset by 60 degrees from each other.
a, Sb, Sc, Sd, Se and Sf are formed. Although only a part is shown in FIG. 18, the stator S
The electric coils CL are wound around the pole portions Sa, Sb, Sc, Sd, Se and Sf, respectively.

Here, the coil CL wound around the pole portions Sa and Sd of the stator S is the first phase, and the coil CL wound around the pole portions Sb and Se of the stator S is the second phase, Poles Sc, S
When the coil wound around f is defined as the third phase, depending on the position of the pole of the rotor R, as shown in FIG.
By sequentially energizing the phase-third phase coil CL,
The rotor R can be continuously driven to rotate clockwise. That is, since the energized pole portion of the stator S constitutes an electromagnet, the pole portion of the rotor R located near the electromagnet is attracted by the electromagnet and rotationally moves. In order to continue the rotation, it is necessary to switch the energization of the coil with the rotational movement of the rotor R. Actually, in the case of this SR motor 1,
Whenever the rotor R rotates 30 degrees, the first coil is energized.
The phase-second phase-third phase may be switched.

The description will be continued with reference to FIG. 1 again. The SR motor 1 has a three-phase coil 1a for driving it,
1b and 1c, and an angle sensor 1d for detecting the rotational position (angle) of the rotor R are provided. Three-phase coils 1a, 1
b and 1c are connected to the drivers 18, 19 and 1A inside the controller ECU, respectively, and are connected to the coil 1
Current sensors 2, 3 and 4 are installed on the signal line connecting a and the driver 18, the signal line connecting the coil 1b and the driver 19, and the signal line connecting the coil 1c and the driver 1A, respectively. ing. These current sensors 2, 3 and 4 are coils 1a, 1b and 1c, respectively.
A voltage proportional to the current actually flowing is output as a current signal S6. In this embodiment, TS2028N94E21 manufactured by Tamagawa Seiki is used as the angle sensor 1d. This angle sensor 1d, as shown in FIG.
It outputs a 10-bit binary signal indicating the absolute value of the angle of 0 to 360 degrees. The minimum resolution of the detection angle is 0.35 degrees.

A CPU is installed inside the controller ECU.
(Microcomputer) 11, input interface 12, current map memory 13a, PWM map memory 1
3b, waveform map memory 13c, power supply circuit 14, current waveform generation circuit 15, addition circuit 16, direction detection circuit 5, D /
An A converter 6, a comparison circuit 7, an output determination circuit 17, drivers 18, 19 and 1A are provided. This controller ECU sequentially calculates the drive speed and drive torque of the SR motor 1 based on the information input from the shift lever, the brake switch, the accelerator switch, and the accelerator opening sensor, and the calculated values are calculated. Based on the results, SR motor 1
The current flowing through each of the coils 1a, 1b and 1c is controlled.

A specific configuration of a part of the circuit shown in FIG. 1 is shown in FIG.
Shown in FIG. 2 shows only a circuit for controlling the energization of the coil 1a of the SR motor 1, but actually the other coil 1b is used.
And a similar circuit for controlling energization of 1c, respectively.

Referring to FIG. 2, one end of the coil 1a is
The other end of the coil 1a is connected via a switching transistor (IGBT) 18b to a low potential line 18f of the power supply via a switching transistor (IGBT) 18a. Further, the emitter of the transistor 18a and the low potential line 18
The diode 18c is connected between the transistor 18f and the transistor 18d, and the diode 18d is connected between the emitter of the transistor 18d and the high potential line 18e. Therefore, if both the transistors 18a and 18b are turned on (conduction state), a current flows between the power supply lines 18e and 18f and the coil 1a, and either one or both are turned off (non-conduction state). The energization of the coil 1a can be stopped.

The output judgment circuit 17 has two AND gates.
17a and 17b and a timing control circuit 17c. The output terminal of the AND gate 17a is connected to the gate terminal of the transistor 18b.
The output terminal of 7b is connected to the input of the timing control circuit 17c. The output of the timing control circuit 17c is
It is connected to the gate terminal of the transistor 18a. The signals S10 and S5 are input to the input terminal of the AND gate 17a, and the signal S7 is input to the input terminal of the AND gate 17b.
1 and S5 are input. The signal S71 is a binary signal output from the analog comparator 7a of the comparison circuit 7. The signal S5 is a binary signal (on / off signal) output from the current waveform generation circuit 15.

At one input terminal of the analog comparator 7a, the result obtained by adding the vibration compensation value and the speed compensation value to the reference current value S4 output from the current waveform generating circuit 15 by the adding circuit 16 is used as a D / A converter. The analog voltage Vr2 converted in 6 is applied, and the voltage (Vs6) of the signal S6 corresponding to the current detected by the current sensor 2 is applied to the other input terminal. The analog comparator 7a outputs the result of comparing the voltages Vr2 and Vs6 as a binary signal S71.

When the signal S5 is at a high level H (energization ON), the ON / OFF of the transistor 18a is controlled based on the binary signal S71 output from the analog comparator 7a. However, the relationship between ON / OFF of the binary signal S71 and ON / OFF of the transistor 18a is not one-to-one, but the timing is adjusted by the timing control circuit 17c.
This will be described in detail later. When the signal S5 is at the high level H (energization ON), the transistor 18b of the driver 18 is turned on / off according to the binary signal S10 input to the AND gate 17a. This binary signal S1
Reference numeral 0 is a signal generated inside the CPU 11, the period of the signal is constant (15 KHz), and the duty is variable. The duty of the signal S10 is the CPU 11
It is changed as necessary by the processing of. In fact, C
The PU 11 refers to the table (see FIG. 13) stored in the PWM map memory 13b connected to the on-duty value based on the rotation speed (rpm) of the motor at that time and the required driving torque. Then, the signal S10 having the duty of this value is output.

That is, in this embodiment, the transistor 18
a and 18b are independent control signals S8
Since the on / off control is independently performed by 1 and S82, the energization control state of the driver 18 includes a state in which the transistors 18a and 18b are both turned on, a state in which both the transistors 18a and 18b are both turned on, and one is turned on and the other is turned off. There are three states, the state and the state.

For example, assuming that the transistor 18b is on at the time of starting energization, the reference level V of the current is
When r2 is switched from 0 to Iref, Vr2> Vs6, so that the transistor 18a is turned on first,
The current flowing through the load gradually increases from 0 with a slope determined by the characteristics (time constant) of the drive circuit and the load. Then, after the current flowing through the load reaches Iref, the transistor 18a repeats OFF → ON → OFF → ON → ... And is controlled so that the maximum value of the current becomes substantially equal to Iref. When the energization is finished, the reference current value is set to I
When switching from ref to 0, Vr2 <Vs6, so the transistor 18a is turned off, and the current flowing through the load gradually decreases to 0 with a slope determined by the characteristics (time constant) of the drive circuit and the load. .

However, in reality, since the control signal S82 applied to the transistor 18b is a pulse signal, there is a period in which the transistor 18b is in the off state even during the rising period at the start of energization. The rising curve of the load current changes. That is, according to the duty of the control signal S82 applied to the transistor 18b, as shown in FIG.
As shown in 2, the rising curve of the load current changes.
Also, even during the falling period at the end of energization, the transistor 1
There is a period in which 8b is in the off state and a period in which the transistor 18b is in the on state, and the falling curve of the load current changes in accordance with the ratio thereof.

Description will be made with reference to FIGS. 23 and 24. As shown in FIG. 23, when both switching means 18a, 18b are turned off from the state where the switching means 18a, 18b are both turned on and the current is flowing to the load 1a, the energy stored in the load 1a is changed by the energy stored in the load 1a. The flowing current flows from the low potential line to the high potential line of the power supply through the diodes D1 and D2. At this time,
Since the potential difference between the terminals of the load 1a is large, the energy is released quickly and the current decay rate is fast. That is, the slope of the fall of the transient current curve is large.

On the other hand, as shown in FIG. 24, when both the switching means 18a and 18b are turned on and the current is flowing through the load 1a, when only one of the switching means 18a is turned off, the other switching means 18b is turned on. Is on, the current flowing by the energy stored in the load 1a is the diode D1, the load 1a,
It passes through the closed loop of the switching means 18b. At this time, since the potential difference between the terminals of the load 1a becomes small,
The energy is released slowly and the current decay rate is also slow. That is, the falling slope of the transient current curve is small.

That is, by adjusting the duty of the control signal S82 applied to the transistor 18b, the waveform of the load current at the time of rising and the waveform at the time of falling can be controlled.

Further, the binary signal S7 output from the comparator 7a
When the chopping control is performed according to No. 1, when the falling speed of the load current is relatively fast, the fluctuation range (amplitude) of the current is large as shown in FIG. 23, and when the falling speed of the current is relatively slow, FIG. As shown in, the fluctuation range of the current becomes small.
By reducing the fluctuation range of the current, in the case of the SR motor, vibration and noise generated during rotation can be significantly reduced.

However, when the falling speed of the current is slow, when the target value (reference level) in the chopping control is changed, a delay in following the current to the target value is likely to occur. The level of the electric current supplied to the motor needs to be changed in accordance with the change of the driving torque. In particular, when driving the SR motor, it is necessary to switch between energization and de-energization of each coil according to the position of the rotor poles, and if there is a delay in following the current with respect to the target value, especially high speed rotation In this case, the rotation torque is significantly reduced.

In this embodiment, the signal S10 is generated on the basis of the rotation speed (rpm) of the motor and the required driving torque.
Since the duty of is automatically adjusted, the number of rotations is high,
Alternatively, when a large driving torque is required, the rise of energization becomes faster, and the delay in following the current with respect to the change in the target value is prevented. Also, when the rotation speed is low or when a large driving torque is not required, the load current rises,
Since the rate of change such as falling is slow, generation of vibration and noise is suppressed. It is difficult to finely adjust the waveform of the current reference level (Vr2) in a short period, but it is easy to adjust the duty of the signal S10.

By the way, according to the comparison result of the comparator 7a, the transistor 18a normally repeats on / off in a short cycle, but if the signal S71 output from the comparator 7a is directly applied to the transistor 18a, the transistor 18a is turned on / off. The OFF period is determined by the characteristics of the energizing circuit, the impedance of the motor coil, and the like, and is also affected by environmental changes such as temperature and humidity. In that case, the on / off frequency of the transistor 18a may be abnormally high. However, when the frequency at which the power is turned on / off increases, the loss in the transistor 18a increases and the amount of heat generation also increases. On the contrary, when the frequency for turning on / off the energization is lower than the upper limit of the human audible frequency, the human being hears the mechanical vibration caused by the switching of the current as noise. Therefore,
The on / off frequency of the transistor 18a is set to a frequency slightly higher than the upper limit of a general human audible frequency (for example, 1
It is desirable to control the frequency to be 5 KHz).

In order to control the on / off frequency of the transistor 18a, the control shown as (a) in FIG. 11 was performed before the device of this example was prototyped. This control will be described. That is, using a synchronization signal with a constant cycle,
Timings t1, t2, t3, ...
A signal S81 for generating and applying it to the transistor 18a
x is switched to the off level every time Vr2 <Vs6,
At each of the timings t1, t2, t3, ...
If Vr2> Vs6, the signal S81x is switched to the on level at that time, but if Vr2 ≦ Vs6, the signal S81x.
Keep off level.

However, in this control ((a) in FIG. 11), Vr2 <V immediately before the timing (t4) of the synchronizing signal.
At s6, the timing (t
In 4), since Vr2 <Vs6, the signal S81x is maintained at the off level. As a result, the period during which the signal S81x is not switched on / off becomes long, and the transistor 18
There were times when the on / off frequency of a fell below the upper limit of human audible frequencies.

Therefore, in this embodiment, control is performed as shown in FIG. 11B by using the further improved timing control circuit 17c. This control is shown in FIG.
This will be described with reference to FIG. Using a synchronization signal with a constant cycle,
Timings t1, t2, t3, ...
・ Generate. The signal FE is switched to a high level H (not on) when Vs6> Vr2, and is set to a low level L at each timing t1, t2, t3, ... Of the synchronizing signal.
Switch to (ON enabled). The condition for switching the signal S81 off is when Vs6> Vr2, and the condition for switching the signal S81 on is that the signal FE can be turned on.
It is when Vs6 ≦ Vr2. According to this control, Vr2 <V immediately before the timing (t4) of the synchronization signal.
The timing of the synchronization signal immediately after that is s6 (t4)
Then, even if Vr2 <Vs6, if Vs6> Vr2 after the signal FE is switched to ON, the signal S at that time.
Since 81 is turned on, the on / off cycle of the signal S81 becomes substantially the same as the cycle of the synchronization signal (reference chopping cycle), and the frequency does not change much. For this reason,
By setting the frequency of the synchronization signal to be slightly higher than the upper limit of the human audible frequency, it is possible to prevent the generation of noise of the audible frequency and also prevent the generation of a large amount of heat.

FIG. 3 shows the actual configuration of the timing control circuit 17c, and FIG. 10 shows an example of the signal waveform of each part in the circuit. In this embodiment, a pulse signal having a frequency of 15 KHz is used as the synchronizing signal CLK15K. The circuit shown in FIG. 3 is a gate circuit 171, 174, 177, 178.
And 179, and D-type flip-flops 172, 17
3, 176 and 17A, and an inverter 175. As shown in FIG. 10, the signal FE is the input signal S71.
Becomes a condition of Vs6> Vr2, it is switched to "ON disabled" and switched to "ON enabled" at the rising timing of the 15 KHz synchronization signal CLK15K. Then, the signal S81 is turned off when the input signal S71 satisfies the condition of Vs6> Vr2, and is turned on when the input signal S71 becomes Vs6 <Vr2 after the “unable to turn on” of the signal FE is released. . Therefore, the control shown as (b) in FIG. 11 is realized by using the timing control circuit 17c.

By the way, in order to make the rotational acceleration of the motor 1 accurately follow the target acceleration, for example, the acceleration force
It is desirable to implement feedback control. In this embodiment, the reference level Vr2 of the current input to the comparator 7a
The current value of the motor 1 is controlled by controlling, but in order to enable the control of the delicate waveform of the current,
An independent current value is assigned to the reference level Vr2 for each minute rotation angle (0.7 degrees) of the motor 1. For this reason,
For example, when the energizing amount of the motor 1 is adjusted in accordance with the change of the rotation speed (rpm) and the required torque, the current values of all angles of the coils of each phase are calculated and those values are calculated. Must be updated and registered in the memory. That is, since it takes a very long time to update the control amount, the response of the control system is very slow. If acceleration feedback control is included in the control system (current waveform generation circuit 15) for adjusting such a large number of current values, a fast response to changes in acceleration cannot be expected.

Therefore, in this embodiment, the acceleration feedback control is carried out as an independent control system other than the generation of the current waveform. That is, as shown in FIG. 2, the acceleration compensation value generated by the acceleration feedback control is added to the output signal S4 of the current waveform generation circuit 15 by the addition circuit 16. Therefore, this acceleration
Since the feedback control system does not include any processing that takes much time, the response speed of the acceleration control system is fast.

In addition circuit 16, in addition to the acceleration compensation value, the vibration compensation value is added. This vibration compensation value is a compensation value for suppressing a minute vibration in the rotation direction of the motor 1. According to an experiment by the present inventor, while the SR motor 1 is rotationally driven in a certain direction, the rotor may temporarily rotate (vibrate) in a direction opposite to the driving direction. It has been confirmed that there is. If the vibration due to such reverse rotation is suppressed, the SR motor 1 is driven more smoothly and the noise is surely reduced.

Therefore, in this embodiment, the reverse rotation of the rotation direction during driving is detected, a vibration compensation value for suppressing the reverse rotation is generated, and the vibration compensation value is input to the adder circuit 16 to compensate the current value. There is. Also in this vibration compensation control system, since it is an independent control system different from the generation of the current waveform, it is possible to sufficiently follow a rapid change (vibration).

In practice, the direction detection circuit 5 shown in FIG. 7 is used to determine the rotation direction CW / CC of the rotor of the SR motor 1 based on the lower 2 bits of the signal output by the angle sensor 1d.
W (forward direction / backward direction) is detected. Referring to FIG. 7, the direction detection circuit 5 includes a D-type flip-flop 5
1, 52 and 58 and gate circuits 53, 54, 55 and 56
And 57. Please refer to the example of the signal waveform of each part of the direction detection circuit 5 shown in FIG. By using the angle sensor 1d and the direction detection circuit 5, it is possible to detect even a slight reverse rotation.

An outline of the operation of the CPU 11 shown in FIG. 1 is shown in FIG.
4 shows. The operation of the CPU 11 will be described with reference to FIG. When the power is turned on, initialization is executed in step 61. That is, after the initialization of the internal memory of the CPU 11 and the mode setting such as the internal timer and the interrupt, the system diagnosis is performed, and if there is no abnormality, the process proceeds to the next processing.

At step 62, shift lever and blur are set.
Read the status of the signals output by the switch, accelerator switch, and accelerator position sensor. Step 62
If there is any change in the state detected in step 6, the process proceeds from step 63 to step 64. When there is no change,
The process proceeds from step 63 to step 65.

In step 64, the target value of the drive torque of the SR motor 1 and the target value of the acceleration are determined based on the various states detected in step 62. For example, when the accelerator opening detected by the accelerator opening sensor increases, the target values of the drive torque and the acceleration also increase. Here, a torque change flag indicating a change in the target torque is set.

In step 65, the detected current SR
Input the rotation speed of the motor 1. The rotation speed is detected by an interrupt process described later. When the rotation speed of the SR motor 1 has changed, the routine proceeds from step 66 to step 68, and when the rotation speed does not change, the routine proceeds to step 67. In step 67, the state of the torque change flag is checked, and when the flag is set, that is, when the target torque changes, the process proceeds to step 68, and when there is no torque change, the process returns to step 62.

In step 68, the PWM map memory 1
3b, the data is input and the duty of the pulse signal (PWM signal) S10 output from the CPU 11 is changed. The pulse signal S10 is constantly output during the driving of the motor 1 and its cycle is fixed at 15 KHz, but the duty is changed according to the state at that time.

That is, the PWM map memory 13b is a read-only memory (ROM) in which various data are registered in advance.
As shown in FIG. 13, a large number of data Pnm (n: corresponding to torque) are associated with various target torques and various rotational speeds (rotational speeds of motors). The numerical value of the column, m: the numerical value of the row corresponding to the rotation speed) are held. For example, since the value indicating 95% of the on-duty is held in the data P34, for example, when the torque is 20 [Nm] and the rotation speed is 500 [rpm], the CPU 11 is -Refer to the contents of P34,
Duty so that the on time of the signal S10 is 95%.
Update the tee.

At the next step 69, the current map memory 13a and the waveform map memory 13c are respectively deselected.
Input data. In this embodiment, the current map memory 13
The a and the waveform map memory 13c are constituted by a read-only memory (ROM) in which various data are registered in advance, and the current map memory 13a holds the data as shown in FIG. The memory 13c holds data as shown in FIG.

That is, the current map memory 13a has a large number of data Cnm (n: columns corresponding to torques) associated with various target torques and various rotational speeds (rotational speeds of motors). , M: numerical value of the line corresponding to the number of revolutions) are stored, and one set of data Cnm includes the energization ON angle, the energization OFF angle, the current upper limit value and the waveform pattern number. ing. For example, the torque is 20 [N · m]
When the number of revolutions is 500 [rpm], the contents of the data C34 are 52.5 degrees, 82.5 degrees, 200 [A] and the waveform pattern number 3. This data C34 indicates the energization information within the range of the rotational position of 0 to 90 degrees.
In the range of 2.5 to 82.5 degrees, the current of the predetermined waveform pattern No. 3 having the upper limit of the current of 200 A is applied, and the range of 0 to 52.5 degrees and 82.5 to 90 degrees. The range of means that the current is cut off. Step 69
Then, Cm selected according to the torque and rotation speed at that time
Input a set of n data.

Further, a set of waveform data corresponding to the waveform pattern number contained in the input Cmn data is read from the waveform map memory 13c. For example, waveform pattern
If the code number is 3, 0, 12, 26,
Input a series of 40, ... This waveform data determines the waveform of the reference value of the current actually passed through the coil. That is, the current reference value is finely adjusted for each angular step of the motor rotor.

In the next step 6A, the energization map data is generated based on the data Cnm and the waveform data input in step 69. That is, a large number of current reference values associated with each angular step of the motor rotor and associated data (details will be described later) are generated. Then, the data of this energization map is used as the current waveform generation circuit 1
Write to the memory (bi-directional memory) inside 5. As will be described later, the current waveform generating circuit 15 has a reference value of 1
Since data for all three phases are automatically generated based on the phase data, only the energization map for a specific one phase is created in step 6A and stored in the memory of the current waveform generation circuit 15. Write.

The CPU 11 executes the above steps 62 to 6A.
Is repeatedly executed. Then, the detected SR mode
If the rotation speed and torque of the motor are constant, step 6
6-67-62, but when the rotation speed changes or the torque changes, steps 68-69-6 are performed.
Since A-6B is executed, the energization map of the current waveform generation circuit 15 is updated.

After the initialization in step 61 is completed, a timer interrupt is generated in the CPU 11 every 4 msec. When this timer interrupt occurs, the CPU 11
The process shown in 5 is executed. This will be described with reference to FIG.

In step 71, the value of the counter TM24 is referred to and it is discriminated whether or not it is a predetermined timing occurring in a cycle of 24 msec. That is, the process proceeds from step 71 to step 81 in FIG. 16 once every 24 msec, and otherwise proceeds from step 71 to step 72.

In step 72, the value of the counter TM8 is referred to and it is discriminated whether or not it is a predetermined timing occurring in a cycle of 8 msec. That is, the process proceeds from step 72 to step 7C once every 8 msec, and otherwise proceeds from step 72 to step 73.

At step 73, it is discriminated whether or not the SR motor 1 is being driven.
4. If not, proceed to step 7A. In step 74, it is identified whether the current driving direction of the SR motor (direction to be driven) is forward rotation / reverse rotation. Then, in the next step 75, it is identified whether the actual rotation direction of the current SR motor is forward rotation / reverse rotation. The actual rotation direction of the SR motor is the direction detection circuit 5
The CPU 11 identifies the actual rotation direction of the SR motor 1 by referring to the binary signal CW / CCW output from the direction detection circuit 5.

In step 76, it is determined whether or not the drive direction of the current SR motor identified in step 74 and the actual rotation direction of the current SR motor identified in step 75 match. If they match, the process proceeds to step 7B and 0 is set to the vibration compensation value. If the drive direction and the actual rotation direction do not match, that is, if the rotor is rotating in the reverse direction with respect to the drive direction due to vibration, the process proceeds to step 78, and a predetermined constant is set as the vibration compensation value. set. In this embodiment, the constant (current value) set to the vibration compensation value in step 78 is +30 [A].
Stipulated in. In the next step 79, the vibration compensation value determined in step 78 or 7B is output and applied to the adding circuit 16. The vibration compensation value is commonly used in the three-phase control system.

In step 7A, the value of the counter TM24 and the value of TM8 are updated (+1). In step 7H, the interrupt timer is reset to generate the next interrupt.

On the other hand, at a predetermined timing that occurs once every 8 msec, first, in step 7C, the states of the shift lever, the brake switch, the accelerator switch, the accelerator opening sensor, etc. are set.
Read through and store the result in internal memory.

In the next step 7E, the rotation speed of the motor is calculated. In this embodiment, since the angle sensor 1d connected to the drive shaft of the SR motor 1 outputs a pulse signal whose period changes according to the rotation speed of the drive shaft, the CPU 1
1 measures the pulse cycle of the signal output by the angle sensor 1d, and detects the rotation speed of the SR motor 1 based on this cycle. The data of the detected rotation speed is stored in the internal memory.

At the next step 7F, the actual acceleration is calculated at different timings based on the data of the plurality of rotational speeds detected at step 7E. For example, by calculating the difference between the latest rotation speed and the rotation speed detected one cycle (8 msec) ago, the speed change per 8 msec, that is, the acceleration can be obtained. The number of rotation speed data and the calculation formula used to obtain the acceleration are changed as necessary. In the next step 7G,
The counter TM8 is cleared.

Next, the operation at a predetermined timing that occurs once every 24 msec will be described with reference to FIG. In step 81, the latest motor rotation speed detected in step 7E is the first threshold value of 5
Compare with 00 [rpm]. And the rotation speed ≧ 500
If so, go to step 84, otherwise go to step 82.
Proceed to. In step 84, the high speed flag is set to 1,
Then, it proceeds to step 85.

In step 82, the latest motor rotation speed detected in step 7E is compared with the second threshold value of 300 [rpm]. And the rotation speed ≦ 3
If 00, proceed to step 86, otherwise proceed to step 83. At step 86, the high speed flag is cleared, and then the routine proceeds to step 87. In step 83,
The high speed flag is checked, and if it is set to 1, then the process proceeds to step 85, and if the high speed flag is cleared, then the process proceeds to step 87.

In step 85, the acceleration compensation value is cleared, and then the process proceeds to step 8A.

In step 87, the target value of acceleration determined in step 64 is compared with the actual acceleration (detection value) obtained in step 7F. And the target value ≧
If it is a detected value, the process proceeds to step 88, otherwise to step 89. That is, in the case of the “first state” shown in FIG. 22, step 88 is executed, and the “second state”
If so, step 89 is executed.

In steps 88 and 89, PI
A D (proportional / integral / derivative) operation is executed to generate an acceleration compensation value. However, the calculation of step 88 and step 89
The content is different from the operation of.

In step 88, the deviation (difference) Ea between the acceleration target value Aref determined in step 64 and the actual acceleration (detection value) Ad obtained in step 7F.
Then, the change in the acceleration deviation Ea per unit time,
That is, the time differential value dE and the integral value IE of the acceleration deviation Ea are calculated, and the acceleration compensation value CP2 is generated based on the following equation (1). In step 89, the next
The acceleration compensation value CP2 is generated based on the equation (2).

CP2 = Kp1 · Ea + Kd1 · dE + Ki1 · IE (1) Kp1, Kd1, Ki1: Constant CP2 = Kp2 · Ea + Kd2 · dE + Ki2 · IE (2) Kp2, Kd2, Ki2: Constant Kp1>Kp1> Kp2, Kd2, Kd2, Kd2 > Ki2 Therefore, if the deviation Ea between the target value Aref and the actual acceleration Ad is Ea
Even if the values are equal, the absolute value of the acceleration compensation value CP2 obtained by the equation (1) becomes larger than the absolute value of the acceleration compensation value CP2 obtained by the equation (2). Since the acceleration compensation value CP2 thus determined becomes the feedback amount of the feedback control system, as shown in FIG.
The feedback amount in the "first state" is larger than the feedback amount in the "second state".

That is, in the case of "target acceleration> actual acceleration" (first state), the amount of current compensation for compensating for the lack of acceleration is relatively large, so the lack of acceleration is sufficiently compensated for to obtain the actual acceleration. It is possible to approach the target acceleration. Also,
If "target acceleration <actual acceleration" (second state),
Since the amount of current compensation for suppressing excessive acceleration is relatively small, it is not possible to make the actual acceleration sufficiently follow the target acceleration, but since the motor current value is larger than the target value, the acceleration performance is improved. It The magnitude of the noise generated by the vibration of the motor in the rotation direction is determined by the magnitude of the deviation between the target acceleration and the actual acceleration. The deviation is very small in the first state and suppressed in the second state. As a result, the average amplitude of the deviation is sufficiently small, so that noise is suppressed.

In step 8A of FIG. 16, the acceleration compensation value is output and input to the adder circuit 16. In the next step 8B, the counter TM24 is cleared.

In the processing shown in FIG. 16, when the rotation speed of the motor is 500 [rpm] or more, or when the rotation speed of the motor exceeds 300 [rpm] and the high speed flag is set. In step 85, the acceleration compensation value is cleared. That is, in that case, the feedback control for compensating the acceleration is not executed. This feedback control is executed when the rotation speed of the motor is 300 [rpm] or less, and when the rotation speed of the motor is 50 [rpm].
This is a case where the speed is less than 0 [rpm] and the high speed flag is cleared.

As described above, in order to obtain the acceleration compensation value used for the feedback control, step 88 or 89.
In the above, a relatively time-consuming calculation process (formula (1) or (2)) including differentiation, integration, multiplication, etc. must be executed. In order to drive the SR motor, the current must be controlled every time the rotational position of the rotor changes by a predetermined angle. In particular, in this embodiment, processing such as current waveform switching and energization on / off angle calculation must be carried out in response to changes in the rotational position of the rotor. However, as the rotation speed of the SR motor becomes faster, the time required for the rotational position to change by a predetermined angle becomes shorter. If the rotation speed of the SR motor is shortened and the rotation speed of the SR motor is increased above a predetermined value, the processing cannot be completed within the allowable time.

The above-mentioned feedback control for acceleration compensation is mainly carried out to suppress the occurrence of vibration. However, in reality, vibration occurs only when the rotation speed is relatively low. Therefore, when the rotation speed of the motor is high, there is no problem even if the feedback control for acceleration compensation is omitted. If the feedback control is omitted, the execution of the PID calculation in steps 88 and 89 can also be omitted. Therefore, the time required for the processing such as the switching of the current waveform and the calculation of the energization on / off angle is long because of the omission. Become. That is, even if the rotation speed of the SR motor is high, it is possible to surely finish the processing such as switching the current waveform and calculating the energization ON / OFF angle within the allowed time.

In the process shown in FIG. 16, when the high speed flag is cleared, the acceleration compensating force is reduced.
If feedback control is executed and the high speed flag is set, the feedback control is omitted. In addition, the threshold value to be referred to for switching the high speed flag is 5
Two of 00 [rpm] and 300 [rpm] are used, and hysteresis is given to the switching of the high speed flag.

FIG. 9 shows a waveform example of the current instruction value S4 output by the current waveform generation circuit 15 of FIG. 1 and the corrected current instruction value S4B. In FIG. 9, CP1 is a vibration compensation value and CP2 is an acceleration compensation value. The vibration compensation value CP1 and the acceleration compensation value CP2 are commonly used for each of the three phases. When the current instruction value S4 is 0, the corrected current instruction value S4B is also set to 0. By adding the vibration compensation value CP1 to the current instruction value S4, the vibration of the SR motor 1 is suppressed, so that the noise is reduced, and the acceleration compensation value CP
By adding 2 to the current instruction value S4, the SR mode is
The response of the acceleration control of the motor 1 is improved.

In this embodiment, the three-phase SR motor 1 is used.
Is driven, the indicated value of the current flowing in each phase coil is set to 3
It is necessary to generate phase components. In this embodiment, it is very difficult to generate the current instruction value S4 because the current instruction value is adjusted for each rotor position (small angle step) to optimize the energization waveform. Moreover, since it is necessary to change the current instruction value each time the rotor makes a minute rotation, the current instruction value S4 must be updated instantaneously. In order to generate such a signal, a large number of current instruction values (energization maps) are registered in the memory in advance, the memory address is associated with the rotor position (angle step), and the rotor position is changed. For each time, the position information may be applied to the address of the memory, and the current instruction value at the position may be read from the memory and given to the current control system. Also,
If three sets of such circuits are installed, it is possible to generate three-phase current instruction values.

However, if three sets of independent current waveform generation circuits are provided for each of the three phases, the required memory capacity will increase and the circuit configuration will inevitably become complicated. Further, the CPU 11 has to rewrite the contents of the memory (energization map) each time the rotation speed of the motor or the required torque changes, but if the memory capacity is large, all the contents are updated. Since a long time is required, the responsiveness of the control system deteriorates.

On the other hand, as shown in FIGS. 9 and 20, the waveforms of the current indication values of the three phases are similar to each other, and only the phases (relative angles) of the waveforms are different from each other. Therefore, 1
It is possible to generate three-phase current instruction values by generating signals whose phases are shifted based on the waveforms of the phase current instruction values.

For example, as shown in FIG.
When a large number of waveform data (hatched portions) associated with each angle step within the range of 90 degrees are held in the memory address associated with each angle step, The second phase waveform data can be obtained by adding +30 degrees to the indicated angle, and the third phase waveform data can be obtained by adding +60 degrees to the indicated angle with respect to the memory address.
You can get the data. Further, by repeating the waveform of the first phase in the range of 0 to 90 degrees, the waveform data in the range of 0 to 360 degrees can be obtained.

That is, only the reference one-phase waveform can be registered in the memory, and the three-phase signal waveform can be generated based on it. By doing so, the capacity of the memory is reduced, the circuit configuration is simplified, and the time required for the process of updating the contents of the memory is shortened.

The actual configuration of the current waveform generation circuit 15 is shown in FIG.
6 and the signal timing of each part is shown in FIG. CPU
The energization map generated by 11 is written in the bidirectional RAM (read / write memory) 49 in the current waveform generation circuit 15.
In this embodiment, the bidirectional RAM 49 has two memory banks. One of these two memory banks is used to read the waveform data, and the other one is used to read the data of the CPU 11. Writing is executed. Therefore, the reading of the waveform data and the writing of the data by the CPU 11 can be executed simultaneously.

The memory bank 1 of the bidirectional RAM 49 is D8.
00H to D886H (H: hexadecimal notation, the same applies below) are assigned to the memory bank 2
Are assigned to memory addresses DC00H to DC86H. The areas in the memory bank 1 are assigned as follows.

D800H to D87FH (128 bytes): Current values for every 0.7 degree of rotation angle (90 degrees: 128 steps) D880H: First phase angle 1 (energization start or end angle) D881H: First phase Angle 2 (energization end or start angle) D882H: second phase angle 1 (energization start or end angle) D883H: second phase angle 2 (energization end or start angle) D884H: third phase angle 1 (energization Start or end angle) D885H: Third phase angle 2 (energization end or start angle) D886H: Waveform convex / concave (energization start → angle 2)
Energization ends in, energization ends in angle 1 → energization starts in angle 2) Memory allocation of memory bank 2 is 400H
It is the same as the memory bank 1 except that it is displaced. Switching between the memory bank 1 and the memory bank 2 is performed by controlling bit 10 (A10) of the address of the bidirectional RAM 49.

The current waveform generation circuit 15 will be described with reference to FIGS. 4 and 6. The 10-bit angle signals RZ0 to RZ9 output by the angle sensor 1d are latched by the latch 41 and applied to one input of the adder 47. Also,
The angle signal RZ0 is applied to the timing pulse generation circuit 42. The timing pulse generation circuit 42 generates clock pulses CLK1A, C based on the 8 MHz clock pulse CLK8M and the angle signal RZ0 generated therein.
LK1B, CLK2A, CLK2B and the latch control signal LATZ are generated.

The 4-bit counter 44 counts the clock pulse CLK2B output from the timing pulse generating circuit 42, and sequentially counts the numerical value in the range of 0 to 15 as the count value CNT.
Is repeatedly output as. The operation of each circuit of the current waveform generation circuit 15 is based on the count value C output from the 4-bit counter 44.
It is determined according to the value of NT. The count value CNT is input to the latch control circuit 45, the angle correction output circuit 46, the address control circuit 48, and the drive signal generation circuit 4C.

The angle correction output circuit 46 is an encoder and outputs the following correction values CPS according to the input count value CNT.

CNT: 0-3, CPS: 0 (0 degree) CNT: 4-7, CPS: 84 (30 degree) CNT: 8-11, CPS: 42 (60 degree) CNT: 12-15, CPS: Therefore, the count value CNT is 0 to 3 at the output of the adder 47.
When, the rotational position of the rotor at that time (angle: RZ0-
RZ9) appears as it is, but the count value CNT is 4 to 7
At the time of, 60 degrees is added (shifted) and the count value CN
When T is 8 to 11, 30 degrees is added (shifted). The adder 47 when the count value CNT is 12 to 15
Output is not used.

The address control circuit 48 determines the following 8-bit value MA07 according to the value of the input count value CNT.
Are output.

CNT: 0, 1, 4, 5, 8, 9, MA07: Output of adder 47 CNT: 2 MA07: 0 CNT: 3 MA07: 1 CNT: 6 MA07: 2 CNT: 7 MA07: 3 CNT: 10 MA07: 4 CNT: 11 MA07: 5 CNT: 12-15 MA07: 6 Further, the latch control circuit 45 inputs the count value CNT.
The 2-bit value MA89 is output as follows according to the value of Further, the memory read signal MRD has a count value C.
It is valid when NT is 0 to 12.

CNT: 0,1,4,5,8,9 MA89: 00H CNT: 2,3,6,7,10-15 MA89: 01H 8-bit value MA07 output from the address control circuit 48.
Is applied to the lower 8 bits of the address of the bidirectional RAM and the 2-bit value MA89 output from the latch control circuit 45.
Are applied to the 8th and 9th bits of the address of the bidirectional RAM. Therefore, the lower 10 of the bidirectional RAM 49
The designated address of the bit is as follows according to the value of the input count value CNT.

CNT: 0, 1, 4, 5, 8, 9, MA07: Output of adder 47 CNT: 2 MA07: 0100H CNT: 3 MA07: 0101H CNT: 6 MA07: 0102H CNT: 7 MA07: 0103H CNT: 10 MA07: 0104H CNT: 11 MA07: 0105H CNT: 12-15 MA07: 0106H That is, the following information is read from the bidirectional RAM 49 in accordance with the input count value CNT.

CNT: 0, 1 DATA: current angle current value (first phase current value) CNT: 2 DATA: first phase angle 1 CNT: 3 DATA: first phase angle 2 CNT: 4, 5 DATA: current angle +30 degree current value (second phase current value) CNT: 6 DATA: second phase angle 1 CNT: 7 DATA: second phase angle 2 CNT: 8, 9 DATA: current Current value at an angle of +60 degrees (current value of the third phase) CNT: 10 DATA: Angle of the third phase 1 CNT: 11 DATA: Angle of the third phase 2 CNT: 12 to 15 DATA: Convex / concave section of waveform The current value of the first phase (DATA: 8 bits) output from the bidirectional RAM 49 when the numerical value CNT is 0 or 1 is latched in the latch 4E in synchronization with the signal PH1C output from the latch control circuit 45. Similarly, the count value CNT is 4,
The current value of the second phase output from the bidirectional RAM 49 at the time of 5 is synchronized with the latch control signal PH2C.
Bidirectional RA when the count value CNT is 8 or 9
The current value of the third phase output from M49 is latched by the latch 4E in synchronization with the latch control signal PH3C. Three sets (three phases) of signals S4 output from the latch 4E are applied to the adder circuit 16 shown in FIG.

On the other hand, the comparator 4B compares the output of the adder 47 with the output of the bidirectional RAM 49. Here, what is actually used is, among the outputs of the bidirectional RAM 49, the first phase angle 1, the first phase angle 2, the second phase angle 1, the second phase angle 2, and the third phase. Angle 1 and phase 3 angle 2
And the convex / concave section of the waveform. That is, the comparator 4
B identifies the magnitude relationship between the current rotor angle (+ shift amount) and the angles 1 and 2 of each phase.

The output of the comparator 4B is generated by the control signal LTCH1 output from the latch control circuit 45 when the count value CNT is 2, 3, 6, 7, 10, 11, 12 and 13, respectively. It is latched and used inside 4C. That is, the output of the comparator 4B when the count value CNT is 2 and 3 is used to generate a binary signal indicating ON / OFF switching of energization for the first phase, and the count value CNT is 6 , 7 is used to generate a binary signal indicating the on / off switching of energization for the second phase, and the comparator when the count value CNT is 10 or 11 is generated. The output of the controller 4B is used to generate a binary signal indicating ON / OFF switching of energization for the third phase. Further, the output of the comparator 4B when the count value CNT is 12 and 13 is used to identify the convex / concave of the waveform and generate a binary signal (S5) indicating ON / OFF of energization.

The 8-bit data output from the adder 47 is in the range of 0 to 127, and the most significant bit is always 0. In addition, the first stored in the bidirectional RAM 49
The phase angle 1, the first phase angle 2, the second phase angle 1, the second phase angle 2, the third phase angle 1 and the third phase angle 2 are also 0 to 1.
It is in the range of 27 and the most significant bit is always 0. On the other hand, regarding the convex / concave section of the waveform held in the bidirectional RAM 49, 255 is assigned to the concave waveform and 0 is assigned to the convex waveform. Therefore, when the comparator 4B compares the convex / concave section of the waveform with the output of the adder 47, the output of the comparator 4B follows only the convex / concave section of the waveform regardless of the output of the adder 47. Determined. Therefore, when the count value CNT is 12 and 13, the drive signal generation circuit 4C determines the convex / concave of the waveform of the three-phase binary signal to be output according to the "convex / concave section of the waveform" information. That is, when the convex / concave section of the waveform is 0: OFF while (current angle) ≦ (first phase angle 1) (first phase angle 1) <(current angle) ≦ (first phase) When the angle 2) of the waveform is on, the binary signal (S5) that is off during the angle 2 of the first phase <(the current angle) is generated, and when the convex / concave section of the waveform is 255: ( On while (current angle) ≤ (angle 1 of the first phase) (off angle 1 of first phase) <(angle of the first phase) ≤ (angle 2 of the first phase) (angle of the first phase) 2) A binary signal (S5) that is ON during <(current angle) is generated. The same applies to the second-phase and third-phase binary signals (S5).

The first phase generated by the drive signal generation circuit 4C,
The second-phase and third-phase binary signals (S5) are output from the drive signal generation circuit 4C and input to the latch 4D. The drive signal generation circuit 4C outputs 3 according to the control signal LAT0 that appears in synchronization with the latch control signal LTCH1 output from the latch control circuit 45 at the timing (CNT: 13) when the states of all binary signals are determined. The two binary signals are latched by the latch 4D and the three-phase two
The value signal S5 is applied to the output determination circuit 17.

In the above embodiment, the first state and the second state
In order to switch the magnitude of the feedback compensation amount CP1 from the above condition, different coefficients Kp1, Kd1, Ki1, Kp2, Kd2 and Ki2 are used in the equations (1) and (2). ing. However, if a process of correcting the acceleration deviation Ea is performed in advance before steps 88 and 89 so that the magnitude is changed between the first state and the second state, steps 88 and 89 are performed.
Then, either one of the equations (1) and (2) can be commonly used in the first state and the second state.

[0113]

According to the first aspect of the invention, compared to the feedback compensation amount in the first state where the target acceleration is larger than the actual acceleration, the feedback in the second state where the target acceleration is smaller than the actual acceleration is compared. Because the amount of compensation is small,
The acceleration characteristics of the electric motor can be improved. Moreover, the vibration feedback of the acceleration can sufficiently suppress the vibration of the rotor and reduce the noise.

Further, according to the third aspect, since the biasing amount (for example, current value) can be finely adjusted for each rotation position (angle) of the electric motor, the rotation of the motor can be made smooth. It is effective to do. Moreover, since there is no fear that the responsiveness of the feedback control is deteriorated, the vibration of the rotor can be effectively suppressed by the feedback control.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a motor drive device of an embodiment.

FIG. 2 is a block diagram showing a detailed configuration of a part of FIG.

FIG. 3 is a block diagram showing a timing control circuit 17c of FIG.

4 is a block diagram showing a configuration of a current waveform generation circuit 15 of FIG.

FIG. 5 is a schematic diagram showing the correspondence between the contents of the memory of the current waveform generation circuit 15 and the generated waveform.

FIG. 6 is a time chart showing the operation of the current waveform generation circuit 15.

7 is a block diagram showing a configuration of a direction detection circuit 5 in FIG.

FIG. 8 is a time chart showing the operation of the direction detection circuit 5.

FIG. 9 is a time chart showing the output of the current waveform generation circuit 15 and the signal after compensation.

10 is a time chart showing the operation of the timing control circuit 17c of FIG.

11 is a time chart showing the operation of the timing control circuit 17c of FIG. 2 and the device before improvement.

FIG. 12 is a time chart showing waveforms of a current reference level Vr2 and a current flowing through a load.

FIG. 13 is a map showing the contents of a PWM map memory 13b.

FIG. 14 is a flowchart showing the processing of the CPU 11.

FIG. 15 is a flowchart showing a part of the timer interrupt processing of the CPU 11.

FIG. 16 is a flowchart showing the remaining part of the timer interrupt process of the CPU 11.

FIG. 17 is a time chart showing an example of changes in acceleration and velocity.

FIG. 18 is a front view showing the basic structure inside the SR motor.

FIG. 19 is a map showing the contents of a current map memory 13a.

FIG. 20 is a time chart showing an example of a basic current waveform when driving the SR motor 1 of the embodiment.

FIG. 21 is a map showing the contents of the waveform map memory 13c.

FIG. 22 is a time chart showing an example of changes in acceleration, current value, and feedback amount.

FIG. 23 is a block diagram showing a path of a current flowing through a conducting circuit.

FIG. 24 is a block diagram showing a path of a current flowing through an energization circuit.

[Explanation of symbols]

1: SR motors 1a, 1b, 1c,
CL: Coil 2, 3, 4: Current sensor 5: Direction detection circuit 6: D / A converter 7: Comparison circuit 7a: Analog comparator 11: CPU 12: Input interface
Phase 13a: Current map memory 13b: PWM map memory 13c: Waveform map memory 14: Power supply circuit 15: Current waveform generation circuit 16: Addition circuit 16a, 16b: Adder 17: Output determination circuit 18, 19, 1A:
Driver 18a, 18b: Transistor (IGBT) 18c, 18d: Diode 18e, 18f: Power supply line 41: Latch 42: Timing pulse generation circuit 43: Counter control circuit 44: 4-bit counter 45: Latch control circuit 46: Angle correction Output circuit 47: Adder 48: Address control circuit 49: Bidirectional RAM 4A: Bank switching circuit 4B: Comparator 4C: Drive signal generation circuit 4D, 4E: Latch R: Rotor S: Stator Ra to Rd, Sa ~
Sf: pole part

Claims (4)

[Claims] 1. An electric motor; a motor energizing means for controlling energization of the electric motor; an acceleration determining means for determining a target acceleration of the electric motor; and an actual acceleration of the electric motor. An acceleration detecting means for detecting the target acceleration determined by the acceleration determining means, an actual acceleration detected by the acceleration detecting means, and a parameter for determining a feedback loop gain.
Negative feedback control means for generating a feedback compensation amount according to the motor and inputting the feedback compensation amount to the motor energizing means so as to suppress the energizing amount of the electric motor. ; And the magnitude relationship between the target acceleration and the actual acceleration is checked to identify a first state in which the target acceleration is larger than the actual acceleration and a second state in which the target acceleration is smaller than the actual acceleration. And the second state, the parameter
An energization control device for an electric motor, comprising: a feedback gain switching means for switching the controller to reduce the feedback compensation amount in the second state compared to the feedback compensation amount in the first state. 2. The acceleration detecting means includes speed detecting means for detecting the rotational speed of the electric motor, and speed difference detecting means for detecting the amount of change in the speed detected by the means per unit time. An energization control device for an electric motor according to claim 1. 3. An electric motor; an energization map means for holding a target value of an urging amount for each rotational position of a rotor of the electric motor.
Rotational position detecting means for detecting the rotational position of the rotor of the electric motor; Energizing amount detecting means for detecting the actual energizing amount of the electric motor; Rotational position detected by the rotating position detecting means Based on the energization amount target value held in the energization map means, the energization of the electric motor is performed according to the energization amount target value and the energization amount detected by the energization amount detection means. A motor urging means for controlling the electric motor; an acceleration deciding means for deciding a target acceleration of the electric motor; and an energization map means for responding to a change in the rotational speed of the electric motor and a change in the target acceleration. Map control means for updating the urging amount target value to be held; acceleration detecting means for detecting the actual acceleration of the electric motor; target acceleration determined by the acceleration determining means and actual acceleration detected by the acceleration detecting means. And the feedback buckle gain Constant to parameters - Fi in response to the other -
Negative feedback control means for generating feedback compensation amount; compensation amount for adding feedback compensation amount generated by the negative feedback control means to the input of the motor energizing means through a path independent of the map control means An adding unit; and a magnitude relationship between the target acceleration and the actual acceleration are checked to identify a first state in which the target acceleration is larger than the actual acceleration and a second state in which the target acceleration is smaller than the actual acceleration, First state and second
The electric mode including the feedback gain switching means for switching the parameters depending on the state, and reducing the feedback compensation amount in the second state as compared with the feedback compensation amount in the first state. Power control device. 4. The acceleration detecting means includes speed detecting means for detecting the rotational speed of the electric motor and speed difference detecting means for detecting the amount of change in the speed detected by the means per unit time. An energization control device for an electric motor according to claim 3.