JP3518901B2 - Driving method and driving device for brushless DC motor - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of driving a brushless DC motor and a driving apparatus therefor, and more particularly to a so-called sensorless DC for detecting and controlling the position of a rotor without using magnetic detection means such as a Hall element. The present invention relates to a brushless motor driving method and a driving device thereof.

[0002]

2. Description of the Related Art Generally, drive control of a brushless DC motor needs to be performed by closely associating the magnetic pole position of a rotor with the position of a winding to be passed. The output torque of the motor is generated by the interaction between the magnetic flux of the magnetic pole of the rotor and the current flowing through the winding of the stator. For this reason, in driving a brushless DC motor, the maximum torque is generated by rotating the motor by passing an electric current through the winding of the phase existing in the vicinity of the maximum magnetic flux generated from the magnetic pole of the rotor. is necessary. Further, the drive control of the brushless DC motor is performed by momentarily switching the phase through which the current should flow according to the rotation of the magnetic pole position of the rotor. The timing of commutation, which is the switching of this phase, is the maximum magnetic pole position. If it deviates significantly from this, the torque generated thereby will decrease, and in the worst case, the motor will step out and stop. Therefore, in the drive control of the brushless motor, it is necessary to detect the magnetic pole position of the rotor by some means and control according to the detected magnetic pole position of the rotor.

On the other hand, as a conventional technique relating to the rotor magnetic pole position detection circuit of this type of sensorless brushless motor, for example, the technique described in Japanese Patent Laid-Open No. 52-80415 is known. This conventional technique delays the induced voltage of the winding of the brushless motor using a bandpass filter, compares the phase delay waveform of the filter with the neutral point of three phases of the filter output voltage, and Flow timing is created, and the inverter circuit that drives the motor is controlled based on this timing to perform commutation.

[0004]

By the way, in the above-mentioned prior art, a phase delay waveform of the induced voltage of the brushless motor is created by using a filter. However, when the filter is used in this way, the phase delay of the filter becomes inaccurate when rotational pulsation occurs in the rotor due to load fluctuation or the like.
Also, under the condition that the motor load is large, when the motor reflux current flows, the time during which the terminal voltage of the stator becomes equal to the DC voltage of the inverter increases, which causes the filter output after filtering the induced voltage of the motor. In some cases, waveform distortion occurs in the voltage, making it impossible to obtain accurate commutation timing.

Further, when the load torque fluctuates repeatedly during one rotation of the motor, if the control is performed so that the output torque of the motor is constant, a fluctuation in speed occurs in the rotation of the motor, which causes a vibration. Destabilize the drive of
In the worst case, the motor may stop.

In order to solve these problems, an inverter circuit is used to change the pulse width to variably control the output AC voltage, so-called Pulse Width Modulation (hereinafter referred to as PWM) control. A driving method is conceivable in which the flow rate is changed every time and the motor torque and the load torque are matched. However, even in this case, if the conventional technique using the filter for determining the commutation timing is used, the difference in the conduction ratio between the conduction periods is reflected in the output voltage of the filter, and the phase delay becomes inaccurate. . Then, due to this inaccuracy, the driving of the motor becomes unstable as in the case described above, and in the worst case, the motor may stop.

As described above, in the prior art in which the commutation timing is determined by using the filter, the correct position of the rotor cannot be detected, and the commutation timing cannot match the magnetic pole position and the speed. There are cases. If the commutation timing cannot be matched with the magnetic pole position and the speed, the generated torque of the motor will eventually decrease, and the motor may step out.

Therefore, in the above-mentioned prior art, the commutation timing caused by the inaccuracy of the phase delay of the filter output is introduced by some method such as introducing another parameter such as a direct current value in addition to the induced voltage of the stator winding. It is necessary to correct the error and determine the commutation timing so that the current flows through the winding of the phase existing near the maximum of the magnetic flux generated by the rotor magnetic field, but in this way the commutation timing is determined. Then, the drive control of the motor becomes complicated.

The present invention has been made in view of the actual situation of the prior art, and the load on the motor is changed, and accordingly, the flow rate is further changed for each flow period so as to change the motor torque. Even when controlling, the magnetic pole position of the rotor is correctly estimated, commutation is performed at an appropriate timing without stopping the motor, the speed fluctuation during one rotation of the motor is reduced, and it is caused by the speed fluctuation. It is an object of the present invention to provide a driving method of a brushless DC motor and a driving apparatus therefor capable of suppressing the vibration and noise caused by the noise.

[0010]

In order to achieve the above object, the first means has a permanent magnet and an armature coil,
In a method of driving a brushless DC motor configured such that one end of each phase of an armature coil is commonly connected, a terminal voltage of a non-energized phase of the armature coil is detected, and a change rate of the detected terminal voltage with respect to time. Is obtained, and the commutation timing is determined based on the obtained change rate.

In this case, the energization of the armature coil is switched by the inverter circuit, and the rotor position at the time of undetection is synchronized with the PWM signal controlling this inverter circuit and the terminals of the non-energized phase of the armature coil. The voltage is detected, extrapolation is performed by the extrapolation based on the change rate of the detected voltage over time, the interpolated value is obtained, and the rotor position at the time when detection is not possible is continuously estimated from this interpolated value to determine the commutation timing. To be done. When determining the commutation timing, the commutation time indicated by the point where the interpolated value by extrapolation with the detected value of the terminal voltage of the non-conducting phase of the adjacent phase is used. The voltage value at the commutation timing estimated based on the interpolation value is reflected at the next rotation. Note that the extrapolation is performed based on the rate of change of the portion of the detected voltage having good linearity.

The detection of the terminal voltage is performed a plurality of times within one PWM cycle, and the number of detections is changed according to the change of the PWM signal conduction ratio and the rotation speed. Further, the calculation by the extrapolation is performed based on the effective induced voltage information of two points, and even when the effective induced voltage cannot be detected even once, the commutation timings are set in advance at each commutation time to the previous time and the current time. It is set by calculating the next commutation time from the commutation time, and is configured so that the motor cannot be driven.

In order to achieve the above object, the second means is
In a drive device for a brushless DC motor that has a permanent magnet and an armature coil, and one end of each phase of the armature coil is commonly connected, select a terminal voltage of the non-energized phase of the armature coil. Non-conducting phase terminal voltage detecting means for analog / digital conversion by means of the analog and digital conversion, and its rate of change and commutation timing are estimated from the terminal voltage detected by this non-conducting phase terminal voltage detecting means, and the electric machine is operated according to the estimated commutation timing And a control means for controlling the drive of the motor by energizing the child coil.

In this case, the non-energized phase terminal voltage detecting means selects a non-energized phase one of the windings of the armature coil and a terminal voltage of the terminal selected by the selector by analog / digital. The control means obtains a rate of change of the detected voltage with respect to time from the detected terminal voltage, and extrapolates the obtained rate of change to obtain induced voltage information. , A controller that estimates the position of the rotor based on this induced voltage information, and determines the flow timing to the armature coil according to the estimated position of the rotor, and the flow timing determined by this control unit. Accordingly, the drive unit controls the energization of the armature coil.

[0015]

According to the first means, the terminal voltage of the one-phase winding (armature coil) of the three-phase winding of the stator armature coil which is not energized is sampled with respect to the permanent magnet of the rotor. The commutation timing is determined based on the rate of change of the terminal voltage with time detected by the sampling, and the motor is driven.

At the time of sampling, for example, a selector and an A / D converter are used to select the terminal voltage of the armature coil in the non-energized phase for A / D conversion, and the A / D converter discretely captures it. The induced voltage information is calculated by extrapolating the detected voltage value based on the rate of change of the detected voltage value with time, and the commutation timing is calculated based on that information to determine the inverter that drives the motor. It drives the driver and finally rotates the motor. At this time,
The captured detected voltage value and time are stored in, for example, a memory, and commutation timing and the like are calculated from the information and the rate of change of the voltage value with time. An inverter circuit, for example, is used for switching for driving the motor.

According to the second means, the non-energized phase terminal voltage detection means selects one of the windings of the armature coil which is not energized, performs A / D conversion, and outputs it to the control means side. In the control means, from the terminal voltage detected by the non-energized phase terminal voltage detection means, obtain the rate of change of the detected terminal voltage with respect to time, and obtain the induced voltage information by extrapolating the obtained rate of change, for example, The position of the rotor is estimated based on this induced voltage information, and the flow timing to the armature coil is determined according to the estimated position of the rotor, and for example, to the armature coil at the timing determined via the drive unit. The motor is driven by controlling the energization of.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a method for driving a brushless DC motor according to the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a block diagram showing a system configuration of a brushless DC motor to which the driving method according to the present invention is applied, FIG. 2 is a terminal voltage detection circuit diagram of the motor, and FIG. 3 is an ideal commutation timing for the motor. 4 is an enlarged view of the waveform of the detection terminal voltage, in which the induced voltage of the motor, the detection terminal voltage, and the processing waveform of the terminal voltage are schematically shown when the flow is controlled by the motor and the motor is rotating at a constant speed. , Fig. 5
FIG. 6 is a waveform diagram showing a processed voltage waveform for explaining the advance and delay of the phase of commutation timing.

As shown in FIG. 1, the brushless DC motor system according to the embodiment includes a DC power source 1, an inverter circuit 2, and a selector 3 for selecting one of the non-energized phases among the three-phase windings. , Converts the analog value of the terminal voltage of the selected terminal into a digital value so that it can be calculated by an electronic computer A
A D / D converter 4, a control unit 5 that determines the rate of change and commutation time from the detected voltage, and outputs a signal to a driver that controls the inverter circuit 2, a driver circuit unit 6 of the inverter circuit 2, It is composed of a stator 7 of a three-phase DC brushless motor in which windings of U, V, and W phases are connected at one end, a rotor 8 having magnetic poles using permanent magnets, and a load 9 connected to the motor. . The driving method of the motor in this system can obtain high output torque.
It is assumed that 0-degree energization type drive is used, the motor control method is so-called PWM control for controlling the DC voltage conduction ratio, and the switching element for chopping is connected to the positive voltage side of the DC power supply 1. . At this time, a PWM frequency of 5 kHz and a motor pole number of 4 were used. The switching element for chopping may be connected to the negative voltage side of the DC power supply 1, and although the inverter circuit 2 is not shown in the figure, three sets of coils are provided to correspond to the windings of each phase. It is provided with a switching element.

In general, in a direct-current motor that is full-wave driven by the inverter circuit 2, effective induced voltage information from the windings is that the induced voltage from one of the three-phase windings which is not energized at all times. However, the induced voltage information is valid during the period when chopping by the inverter circuit 2 is not performed. Therefore, the selector 3 has a switching function of switching the terminals of the detected phases according to the conduction mode, and the A / D converter 4 converts the analog voltage value into a digital value in synchronization with the PWM signal. So-called A /
It has a D conversion function.

The control unit 5 receives the detected value of the induced voltage and calculates the rate of change of the voltage with respect to time, that is, the slope of the voltage based on the detected value, and the time from the slope to the target voltage value. And a function of outputting a signal that controls the driver circuit 6 that controls the switching element of the inverter circuit 2, and a function of storing the above-described detected value and inclination.

The driver circuit section 6 has a function of driving a switching element forming the inverter circuit 2 according to a signal from the control section 5. In this embodiment, the control unit 5 and the driver circuit unit 6 of the switching element of the inverter circuit 2 are formed by different circuits.
A microcomputer in which these functions are integrated may be used as the control unit.

The stator 7 constituting the motor is composed of three-phase windings U, V, W which are magnetically poled by passing an electric current, and one ends of these windings are connected to each other. . The rotor 8 is a permanent magnet having a magnetic pole, and rotates according to the appropriate change in the magnetic pole position of the stator 7 described above. Further, as can be seen from FIG. 2 showing the terminal voltage detection circuit of the motor, the U phase of the stator 7 of the motor, the V
When the induced voltage is detected from each of the W-phase and W-phase terminals, the induced voltage is divided by using the operational amplifier 10, and one phase detected by the selector 3 is selected.
Optimal induced voltage information can be sent to the / D converter 4. Next, in the brushless DC motor configured as described above, it is assumed that each winding of the stator 7 is controlled to flow at an ideal commutation timing and the motor is rotating at a constant speed. This will be described with reference to FIG.

FIG. 3A is a waveform diagram showing the induced voltage waveform of each phase which occurs as a physical phenomenon in this state.
FIG. 3B is a waveform diagram showing the terminal voltage waveforms of the windings of each phase detected in full-wave drive, and FIG.
It is a waveform diagram which shows the signal waveform which can be extrapolated and interpolated from the amount of change with time of a detected value about the point which cannot be detected in synchronization with a WM signal.

As shown in FIG. 3, when the motor commutates at an ideal commutation timing and rotates at a constant speed, a reverse voltage as shown in FIG. 3A is induced as a physical phenomenon. To be done. This induced voltage is induced by the magnetic pole of the rotor, its Peak To Peak value is proportional to the rotation speed of the motor, and the transient value is determined in correlation with the magnetic pole position of the rotor. Usually, the commutation timing of the motor can be determined based on this induced voltage information.

That is, the ideal commutation timing (maximum output can be obtained) of the motor is the points tp1 and tp2 shown in the drawing where the induced voltages of the adjacent phases intersect, and in this embodiment, The voltage at this time is obtained as the maximum value e (tp1) of the voltage which is an upward convex inflection point and the minimum value e (tp2) of the voltage which is a downward convex inflection point. When the motor is rotating at a constant speed, the maximum value and the minimum value of the voltage at that time are almost equal in any adjacent phases.

On the other hand, what can be detected as an actual system is the terminal voltage of the stator winding of the motor, which has a waveform as shown in FIG. 3 (b). The drive voltage and the induced voltage applied to rotate the motor are mixed. Then, in the 120-degree conduction type drive of the motor, the effective induced voltage information is the period Ts shown in FIG.
Only one of the phases. Further, even in the period Ts of effective induced voltage information, the potential is fixed to the positive or negative value of the DC voltage in the section where the current flows through the freewheeling diode during commutation, and the inverter circuit is PWM.
Since chopping is being performed when a signal is being sent, it does not become continuous as shown in FIG. For this reason, it is necessary to detect the induced voltage by some method and perform arithmetic processing based on the information to determine the commutation timing.

Therefore, in this embodiment, effective induced voltage information is obtained from only one phase among the three phases during the period Ts shown in FIG. 3B in the 120-degree conduction type drive as described above. Therefore, the selector 3 switches and selects one of the three phase windings of the motor that is not energized according to the current flow mode, and the selected waveform is shown in the enlarged view of the detected waveform in FIG. As described above, no current is flowing through the freewheeling diode in the inverter circuit 2, and the terminal voltages e (t1) and e (t2) at the time of chopping off are sampled, but are discrete, but The induced voltage information is directly picked up from the voltage.

Incidentally, the term “discrete” here means that the induced voltage information obtained from the motor is in the ON period of the PWM signal, so that when the PWM signal variably changes, A /
It is also necessary to variably change the sampling point for D conversion. That is, even if the A / D converter is continuously operated, effective information can be obtained only during the ON period of the PWM signal, so that the sampling is necessarily discrete. Conversely, if effective induced voltage information is obtained even during the OFF period of the PWM signal, continuous sampling may be performed instead of discrete sampling.

Further, in this embodiment, the rate of change Δv / Δt of the voltage with respect to time is calculated from the induced voltage information discretely obtained by the control unit 5, and the induced voltage between discrete values (the broken line portion in the figure) is calculated. Obtained by extrapolation and continuously connected for three phases to obtain a processed waveform having a triangular wave signal waveform as shown in FIG.

In the state where the flow of current to the windings of each phase is controlled at the ideal commutation timing and the motor is rotating at a constant speed, as shown in FIG. The voltage at the point where the voltage intersects (the maximum value of the voltage that is the upward convex inflection point, the minimum value that is the downward convex inflection point) becomes almost constant in each adjacent phase. Of the processed waveform from the voltage value at each inflection point of this signal
And emin can be determined. From this, in order to find the optimum commutation timing, the induced voltage in each phase is detected, and the time at which the induced voltage obtained from the rate of change with respect to time is emax and emin is obtained. When the phase of the commutation timing is delayed or advanced, as shown in (b) and (c) of FIG. 5, the voltage value e (t '41) and e (t'42) and the voltage values e (t41) and e (t42) at the time of the previous rotation are taken, and at the time of the previous commutation based on the current or previous rate of change of the induced voltage. By estimating and comparing the optimum values of the voltage value e (tpb) at t41 and t42, the delay of the previous commutation time and the advance time Δtbd are estimated, and the next commutation point without accumulating the commutation time error. Is corrected in. Then, by repeating such correction, it is possible to finally bring the commutation timing closer to the optimum commutation time. In addition, since the induced voltage information depends on the rotational speed and the magnetic pole position of the rotor 8, even when the motor causes rotational pulsation due to the above-described correction, the induced voltage also changes due to this rotational pulsation, so that proper commutation is always performed. The timing can be estimated.

The above-mentioned processing for obtaining the commutation time will be described in more detail below. Now the load on the motor 9
However, it is assumed that the motor is rotated at a constant speed with a constant load having no periodic load fluctuation. In this case, a voltage waveform as shown in FIG. 4 is detected as the induced voltage.
Sampling for obtaining the induced voltage information from this voltage waveform validates the data only during the period when the PWM output of the inverter circuit 2 is ON and the current does not flow through the freewheeling diode during commutation. Done.

The data of the invalid period is calculated by extrapolation from the detected voltage value and time using the change rate Δv / Δt. In this way, by obtaining the voltage value in the ineffective period by extrapolation interpolation, it is possible to continuously estimate the rotor position for each conduction period from the time or voltage obtained as a result.

Further, the commutation time is estimated from the rate of change in induced voltage Δv / Δt obtained on the basis of the present time, which is the maximum value which is the upwardly convex inflection point obtained at the time of one revolution last time, or The time until the minimum voltage, which is a downward convex inflection point, is calculated as the reference time tbm, and the commutation phase advance and delay time tbd are estimated based on the information of the previous rotation flow point, This is performed by a method of correcting the advance and delay of the previous commutation time with respect to the reference time tbm by the time tbd.

In practice, the voltage value in the detectable region is treated as being represented by an approximate expression of a linear function with respect to time, and the rate of change K = Δv / Δt with respect to time is the voltage value sampled as shown in FIG. It is obtained from e (t1) and e (t2) by the equation (1).

K = Δv / Δt = {e (t2) −e (t1)} / (t2−t1) (1) Next, based on the current time tn, the value obtained at the previous one rotation is The next rotational flow reference time tbm, which is the time until the voltage reaches the maximum value which is a convex inflection point, is calculated by the equation (2) based on the change rate K and the voltage value e (tn) at the current time. At this time, it is desirable that e (tn) be a voltage value at a position of 1/5 to 4/5 where the most linear approximation is effective in the time width from emax to emin.

In the formula (2), the maximum value is described, but the same applies to the minimum value. Then, as shown in FIG. 5A, if the phase of the commutation timing advances and there is no delay, the voltage values of the adjacent phases match at the commutation time, and the rotor is continuously rotated based on the change rate K described above. The position can be estimated and the commutation timing can be determined.

Tbm = {emax-e (tn)} / K (2) Further, as shown in FIGS. 5B and 5C, the commutation phase generated at the previous commutation is delayed, If there is a lead, the correction time tbd is the time at which the voltage value at the previous rotational flow point is estimated from the current change rate K and the change rate Kb obtained from the previous rotational flow time, and the voltage values match. That is, first, the voltage e (t'41) at the previous commutation time, which is the period during which the current flows through the freewheeling diode, is estimated by the equation (3).

E (t′41) = e (tn) − (tn−t41) · K (3) Further, the pre-rotational flow previously obtained from the change rate Kb used in the previous commutation time calculation. Estimated voltage at time e (t4
1) and the half point is assumed to be the voltage value when optimal commutation is performed, and the voltage value and the current rate of change K
The previous phase time tbd (n-1) is added to the time obtained from the voltage e (t'41) obtained from the above, and the next rotational flow correction time tbd (n) is calculated as shown in equation (4). It is the phase delay time at the previous rotating flow point. The voltage value assumed at this time is reflected in the next rotation flow as the maximum voltage value at the ideal commutation timing.

Tbd (n) = {e (t'41) -e (tpb)} / K + tbd (n-1) (4) Finally, the time tb until the next rotational flow time is The rotation flow reference time tbm is the time delayed by the phase delay time tbd according to the equation (5).

Tb = tbm + tbd (n) (5) Although the above description is for the case of commutation phase delay shown in FIG. 5B, in the case of phase advance as shown in FIG. 5C, the rate of change is The same applies when is negative.

In the above description, the commutation time is obtained from the detected change rate of the induced voltage of the motor, and the commutation is performed by managing the time at that time. However, the change rate of the induced voltage of the motor is Since it has voltage and time information, it compares the detected voltage or extrapolated voltage with the voltage value estimated to be the commutation timing, and when it matches or exceeds that voltage. A method of controlling the drive of the motor by controlling with a voltage that causes commutation is also effective.

That is, the rate of change of voltage with time Δv / Δ
When t is used, since both voltage and time information are stored, the rotor position and commutation timing can be controlled by both voltage and time.

On the other hand, the PWM pulse number S (pieces) in one flow period (electrical angle within 60 degrees) is the rotation speed N (rpm), 3
It is determined by the equation (6) by the number of poles P of the phase motor and the PWM cycle Tpwm (sec).

S = 60 · Tpwm / 3 · P · N (6) FIG. 6 shows changes in the number of PWM pulses with respect to the rotation speed when the PWM cycle is 200 (μs). Also, the number of PWM pulses within an electrical angle of 60 degrees depends on the magnitude of the number of revolutions, and as shown in FIG. 6, since the time for an electrical angle of 60 degrees becomes longer in the low speed rotation range, the number of PWM pulses increases. At the same time, the number of times the sampling of the induced voltage is increased. However, in the high-speed rotation range, the time for an electrical angle of 60 degrees becomes shorter,
The number of PWM pulses decreases, and the number of samplings of the induced voltage also decreases. Further, the induced voltage information necessary for estimating the commutation timing is T1 shown in FIGS. 7 (a) and 8 (a).
(The ON period of the PWM signal), and the freewheeling diode has no reflux current flowing therethrough. FIG. 7B is a diagram showing sampling points S1 and S2 in the case where the induced voltage is detected once for each PWM cycle (hereinafter referred to as one-point sampling). Information can be obtained enough to estimate the commutation timing, but in the high-speed rotation range,
As shown in FIG. 6, since the number of PWM pulses within an electrical angle of 60 degrees decreases as the rotation speed increases, the number of times the induced voltage is sampled also decreases.

For example, a high speed rotation range, especially 10,000 rpm
The number of PWM pulses in the equation is 2.5 according to the equation (6), and if the period during which the current is flowing in the freewheeling diode is excluded, the number of obtained PWM information is 2.5 or less. . Therefore, if the number of PWM pulses within an electrical angle of 60 degrees is one and the number of samplings of the induced voltage within one PWM cycle is one (hereinafter, 1
Point sampling), the rate of change of voltage with time Δ
It becomes impossible to calculate v / Δt, it becomes difficult to estimate the optimum commutation timing, and it becomes impossible to drive the motor in the high speed rotation range.

From the above, it is necessary to increase the number of samplings of the induced voltage within the electrical angle of 60 degrees in order to enable the motor drive in the high speed rotation range. Therefore, PWM
By detecting the induced voltage in one cycle a plurality of times (hereinafter referred to as a plurality of samplings), it becomes possible to drive the motor in the high speed rotation range. For example, as shown in FIG.
When the number of detections of the induced voltage in one cycle is set to 2 (hereinafter, 2-point sampling), the amount of information on the induced voltage in each PWM cycle is doubled, so that the calculation accuracy for determining the commutation timing is high. Rises, P within an electrical angle of 60 degrees
Even if the WM pulse is one pulse, the induced voltage can be detected twice, so that the commutation timing can be estimated from the sampled two points S1 'and S2', and the motor can be driven in the high-speed rotation range.

Therefore, as shown in FIG. 7B, T1 <T
When 2 (T1 / Tpwm <50%), one point sampling is performed for each PWM cycle, and as shown in FIG.
When ≧ T2 (T1 / Tprm ≧ 50%), the information for performing extrapolation interpolation is optimally detected by changing the sampling number to two-point sampling according to the change of the rotation speed and PWM. be able to. The procedure for detecting the induced voltage under the above conditions will be described with reference to the flowchart in FIG.

In this procedure, first, one phase is selected by the selector 3 from the three phases U, V and W in order to detect the induced voltage (step 1). Further, the detection phase may be selected in the microcomputer of the control unit 5 without using the selector 3. Next, in order to optimally sample the induced voltage information required for extrapolation interpolation calculation, the start timing of A / D conversion is set during the ON period of PWM (step 2). Then, the number of rotations of the motor is checked (step 3), and the number of times of sampling is changed according to the following conditions.

That is, when the rotation speed is less than 3000 rpm (NO in step 3), one-point sampling is performed regardless of the PWM conduction ratio (step 7).

When the rotation speed is 3000 rpm or more (YES in step 3), the sampling frequency is changed according to the PWM conduction ratio. In this case, the PWM conduction ratio of 50% is used as a reference, and if it is equal to or higher than this reference, in other words, if the conduction ratio is 50% or more (YES in step 4), two-point sampling is performed in step S5, and the conduction ratio If it is 50% or less (NO in step 4), the optimum commutation timing can be determined by performing one-point sampling in step 7. Further, in the above-mentioned detection method, when the conditions such as the PWM cycle and the number of poles of the motor are changed, the rotation speed in step 3 and the PWM conduction ratio in step 4 are different.

The commutation timing needs to be estimated from the point where the induced voltage has the best linearity. Therefore,
As a method of estimating the commutation timing, extrapolation interpolation calculation is performed at the time when two effective induced voltages are detected, and when there are a plurality of detection times, the commutation timing is changed each time the calculation is performed (step 6). FIG. 10 shows that when the commutation timing is estimated, extrapolation interpolation calculation is performed from the detected induced voltages at two points, the commutation timing is changed each time the calculation is performed, and the commutation timing to be finally obtained (ideal FIG. 6 is a diagram showing a state of approaching a different commutation timing).

As shown in FIG. 10, effective induced voltage information is after t1 which is not affected by the freewheeling diode's return current or the like. Then, as one method of extrapolation interpolation calculation, the gradient is obtained from the induced voltages at the two points detected at t1 and t2 shown in the figure, and the commutation timing tm1 is estimated from the intersection of y1 and emax. Next, the slope is obtained from the induced voltages at the two points detected from t2 and t3, and the commutation timing tm2
Is estimated, and tm3, tm4, ... Are sequentially calculated. Then, by comparing the commutation timings calculated up to that time every time the extrapolation interpolation calculation is performed, the point at which the commutation timing converges most can be determined as the ideal commutation timing. In this embodiment, 2 from emin to emax
The calculation is terminated at the point of / 3, and this point is used as the commutation timing. Further, it is not necessary to detect the induced voltage after tn + 1 until the next commutation, and the processing time of software can be greatly reduced.

The procedure for calculating such commutation timing will be described with reference to the flowchart of FIG. In this processing procedure, first, the induced voltage is detected from a point that is not affected by the reflux current or the like (step 11), and when the induced voltage at two points is detected, extrapolation interpolation calculation is performed to obtain the slope (step 11). 12). Next, the commutation timing is calculated from the obtained inclination (step 13), and if the induced voltage information is only two points, the commutation timing is estimated from this information and the commutation is performed. When a plurality of pieces of induced voltage information are obtained, the commutation timing is changed each time the extrapolation interpolation calculation is performed. And the detection point of the induced voltage is from emin to emax
When you reach the point of 2/3 of
ES), the calculation of commutation timing is completed (step 1
5) After that, the induced voltage is not detected until the next commutation. Eventually, the commutation will occur at the commutation timing when the extrapolation interpolation calculation is finally performed within the electrical angle of 60 degrees.

As described above, by calculating the next commutation timing in real time immediately after the commutation, the commutation delay due to the calculation delay can be prevented, and finally the ideal commutation timing can be approximated. be able to.

On the other hand, since the number of samplings of the induced voltage is reduced in the high speed rotation range, the effective induced voltage is 1
If the number of times cannot be detected, the commutation timing cannot be estimated and the motor cannot be driven. Therefore,
If the induced voltage cannot be detected, it is necessary to estimate the current commutation timing from the previous commutation information.

FIG. 12 shows the rate of change Δv of voltage with time.
It is the figure which showed the relationship of the software processing performed every time at the time of commutation in order to output the PWM signal and the triangular wave signal waveform which carried out the extrapolation interpolation from / deltat and connected continuously three phases. As shown in FIG. 12, when the soft processes S1, S2, S3 ... Sn are performed during commutation, the current commutation time tS2 and the previous commutation time tS2 within the t period (soft processing time of S2) shown in the figure. The next commutation time tS3 is estimated from the commutation time tS1 by the equation (7).

TS3 = tS2 + (tS2-tS1) (7) That is, before performing the extrapolation interpolation calculation, tSn + 1 (next commutation time) is set in advance at tSn (commutation) as shown in the figure. By presuming, even if the effective induced voltage cannot be detected even once, the commutation can be performed and the motor can be driven.

[0060]

As described above, according to the present invention, it is possible to accurately estimate the position of the rotor even when the flow rate differs for each flow period. This makes it possible to control the motor torque so that it matches the load torque by changing the current flow rate for each current flow period in response to periodic load fluctuations. The noise can be suppressed. Furthermore, even when the number of PWM pulses decreases during high-speed rotation, it is possible to estimate the optimum commutation timing by setting the number of detections within one PWM cycle multiple times, and from the low-speed rotation range to the high-speed rotation range. The motor can be driven accurately.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of a drive device for a brushless DC motor according to the present invention.

FIG. 2 is a circuit diagram showing a terminal voltage detection circuit for detecting a terminal voltage of an armature coil of a motor.

FIG. 3 schematically shows waveforms of an induced voltage, a detection terminal voltage, and a processing voltage of the motor in a state where the commutation is controlled at an ideal commutation timing with respect to the motor and the motor is rotating at a constant speed. FIG.

FIG. 4 is a waveform diagram showing an enlarged detection terminal voltage waveform.

FIG. 5 is a waveform diagram showing a processing voltage waveform for explaining the advance and delay of the phase of commutation timing.

FIG. 6 is a diagram showing the number of times PWM pulses are detected within an electrical angle of 60 degrees.

FIG. 7 is an explanatory diagram showing the relationship between time and voltage when one point of the induced voltage is sampled during the ON period of the PWM signal.

FIG. 8 is an explanatory diagram showing a relationship between time and voltage when the induced voltage is sampled at two points during the ON period of the PWM signal.

FIG. 9 is a flowchart showing a procedure for detecting an induced voltage.

FIG. 10 is an explanatory diagram showing a state in which the commutation timing is estimated from the induced voltage information at two points, the commutation timing is changed every time the calculation is performed, and the ideal commutation timing is approached.

FIG. 11 is a flowchart showing a procedure of calculating commutation timing.

FIG. 12 is an explanatory diagram showing a relationship between a waveform when extrapolation is performed from a detected induced voltage and software processing performed during commutation.

[Explanation of symbols]

1 DC power supply 2 Inverter circuit 3 selector 4 A / D converter 5 control unit 6 Driver circuit 7 Stator 8 rotor 9 load

─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Motoo Futami 7-1-1 Omika-cho, Hitachi-shi, Ibaraki Hitachi Ltd. Hitachi Research Laboratory (72) Inventor Yasuo Notohara 7-chome, Omika-cho, Hitachi-shi, Ibaraki No. 1 No. 1 in Hitachi Research Laboratory, Hitachi, Ltd. (56) Reference Japanese Patent Laid-Open No. 7-123773 (JP, A) Japanese Patent Laid-Open No. 5-15190 (JP, A) Japanese Patent Laid-Open No. 5-236787 (JP, A) Kaihei 5-184190 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) H02P 6/08

Claims (7)

(57) [Claims] 1. A brushless DC motor having a permanent magnet and an armature coil, wherein one end of each phase of the armature coil is connected in common , wherein a non-energized phase of the armature coil is provided. detecting a terminal voltage, determine the rate of change with respect to time of the detected terminal voltage, and drives the motor to determine the commutation timing based on the obtained rate of change, this time, before
Switching the energization to the armature coil with an inverter circuit
The PWM signal that controls this inverter circuit.
The terminal voltage is detected in anticipation of the
The extrapolation value is calculated according to the change rate
Continuously estimate the rotor position when it cannot be detected from the interpolated value,
Extrapolation supplement to the detected value of the terminal voltage of each adjacent phase
The commutation time is defined as the point at which the interpolated values of the
Determine the timing, and based on the interpolated value by the extrapolation
The voltage value at the estimated commutation timing is
The extrapolation interpolation is reflected in the detected voltage straight line.
In the drive method for a brushless DC motor, which is performed based on the rate of change of a good portion, the detection of the terminal voltage is performed a plurality of times within one PWM cycle. Brushless DC motor driving method. 2. An end which is performed a plurality of times within one PWM period.
The number of times the slave voltage is detected depends on the PWM signal conduction ratio and rotation speed.
The method for driving a brushless DC motor according to claim 1, wherein the method is changed according to a change . 3. The calculation by the extrapolation is effective at two points.
Claim 1, characterized in that by such induced voltage information
A method for driving the described brushless DC motor. 4. The commutation time from the previous time and this time is changed to the next for each commutation time.
Calculate the commutation time in advance and estimate the commutation timing
, The effective terminals required to determine the commutation timing
If the voltage cannot be detected even once, the estimated transfer
The driving method of a brushless DC motor according to claim 1, wherein the driving the motor based on the flow time. 5. An electric machine having a permanent magnet and an armature coil.
One end of each phase of the child coil is commonly connected.
In a drive device for a rackless DC motor, the armature coil
Select the terminal voltage of the non-energized phase of analog
Non-conducting phase terminal voltage detection means for digital conversion and this non-conducting state
From the terminal voltage detected by the phase terminal voltage detection means,
Change rate and commutation timing of the
Current to the armature coil according to the current timing.
Features and be Lube Rashiresu DC motor driving device that a control means for driving and controlling over data. 6. The non-energized phase terminal voltage detection means is an electric machine.
Selection to select one of the non-energized phases of the winding of the child coil
Connector and the terminal voltage of the terminal selected by this selector.
Analog / digital for analog / digital conversion of pressure
Brushless DC motor driving device according to claim 5, wherein Rukoto such from the Le converter. 7. The control means detects whether the detected terminal voltage is detected or not.
Then, the rate of change of the detection voltage with time is calculated from
The rate of change is extrapolated to obtain the induced voltage state.
The rotor position is estimated based on the pressure information, and the estimated
The timing of the flow to the armature coil depending on the position of the
Control unit that determines the flow rate and the flow rate determined by this control unit.
Drive that controls the energization of the armature coil according to the imming
Parts and Rukoto brushless DC motor driving device according to claim 5, wherein a from.