JPH02299496A - Driving method for brushless dc motor - Google Patents

Abstract

PURPOSE:To improve speed-torque characteristic and efficiency considerably by performing control such that the phase difference between the current flowing through a winding and the voltage induced in the winding will be approximately zero at all times. CONSTITUTION:Clock signals CLK1, CLK2 for detecting the rotary position of a DC brushless motor and signals U, V, W for detecting the positions of a permanent magnet in U, V and W phases are inputted through a digital input section. A current phase delay detecting section 3 calculates SIGMAiu*sintheta and SIGMAiu*costheta for every clock timing based on these signals and U phase current iu, then the calculated values are integrated by electrical angle of 2pi and employed for calculation of delay angle psi between induced power (e) and winding current iu. PI control section 5 leads the winding application timing by lambda deg.so that the delay angle psi is zero. When lead angle control is carried out similarly for other phases, the power factor approaches to 1.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a driving method for improving the speed/torque characteristics and efficiency of a brushless DC motor.

(Prior art) As described in page 4 of "rDC Brushless Smoke and Control Circuit" published by Sogo Denshi Publishing Co., Ltd. (authored by Kinji Tanikoshi), a brushless DC motor uses a mechanical rectification mechanism. This is a motor that eliminates problems such as noise, wear, and maintenance of conventional DC motors by replacing it with a rotating magnetic field detection element and a semiconductor switching element, and is mainly used in AV and FA.
It is often used in fields with a relatively narrow range of speed and small capacity, such as OA equipment. This is because motor efficiency, heat generation, etc. are not so important, and speed-torque characteristics are not required over a wide range.

In order to make it easier to understand the major differences between Amate capacity, medium capacity, and large capacity, here is a timing chart of the single-phase induced electromotive force e, the voltage V applied to the motor winding, and the current i flowing through the motor winding. Shown in Figure 1.

(a) is a typical example of small capacity, and (b) is a typical example of medium and large capacity.

The sensor that detects the magnetic pole position (for example, a magnetic detection element such as a Hall generator, or a photodetection element that uses light emission or reflection) is placed at an appropriate position so as to synchronize with the timing of the induced electromotive force e at the time of small capacity in (a). The applied voltage (2) is the one whose timing and direction to apply to the motor windings are determined according to the rotational movement and polarity of the magnetic poles. This applied voltage V causes a winding current i to flow. This also applies to (b).

Compared to the winding current i of medium and large capacities (b), the winding current i of the small capacity (a) flows at almost the same timing as the applied voltage (2), but there is a delay time in (b). (a) has a small capacity, so the motor winding itself is thin, the total motor winding resistance is very large, 5 to 10 Ω, and the winding time constant determined by the motor winding inductance is also very small. many.

On the other hand, in (b), since a large capacity motor generates a relatively large torque, it is necessary to flow a large current, so the windings are thicker than those of a small capacity motor, and as a result, the winding resistance becomes smaller, and the number of turns increases. The winding time constant also changes in relation to the winding inductance, which is determined by the motor volume, etc., and generally becomes larger, so that the current i flows much later than the timing of the winding applied voltage (2).

In other words, the motor output is determined by the power factor determined by the induced electric power e, the fundamental wave component of the phase current i, and the phase difference between the two. Therefore, in (a), since the induced electromotive force e and the winding current i are almost in phase, the power factor becomes 4=:l, indicating that the efficiency is very good. Furthermore, in (b), the portion (time) where the product of induced power e and winding current i is negative (rather than positive) is longer, and the period of reactive power is longer than in (a). This tendency often increases with larger outputs and larger time constants. Therefore, in method (b), the output is small compared to the input energy, and the winding is heated and the temperature rises, resulting in a lot of waste.

The heat generated by the windings causes the entire inside of the motor to rise in temperature, sometimes reaching temperatures as high as 1
It is not uncommon for the temperature to exceed 00°C, and it is possible that the temperature exceeds the heat resistance limit of the magnetic pole position detection element. In addition, in the case of brushless DC motors, the voltage applied to the windings corresponding to the magnetic pole position is also pulsed, so the winding current contains many harmonics, which also causes heat generation. .

As mentioned above, small-capacity brushless DC motors that require less reactive power section reduction and harmonic current removal do not take these measures, and are oriented toward high rotation and low torque. Torque types are extremely rare. Therefore, as a method for improving these, we have already applied for the patent application 1986-75796 "Method of driving a brushless DC motor".
However, this method also has the following problems.

(1) When creating a sin curve or cos curve for every Δθ (for example, 3 @) to detect the phase difference between the induced power and the winding current, the N-pole to S-pole section is created in the circumferential direction of the permanent magnet. This was done using a complicated method in which a Hall generator or optical or magnetic sensor was installed, and a large number of amplifiers were installed, and the distribution was adjusted using resistors to obtain the sin and cos values every 3 degrees, and then addition was made. , which is time consuming, difficult to coordinate, and costly.

(2) In a motor with a small diameter, the diameter of the permanent magnet is also small, and the length between the N and S poles in the circumferential direction is short. There is a limit to how many wires can be installed, and the number of wiring increases, and since it is close to a rotating body, it is highly dangerous in practice.

(3) If a magnetic sensor such as a Hall generator is used as the sensor in (2) above, the influence of a large current magnetic field will cause the Hall generator to malfunction, making it difficult to detect the magnet position, and causing sin , the cos curve becomes abnormal.

The current situation is that there are problems as described above.

[Problem to be solved by the invention]

As mentioned above, if the voltage applied to the winding is supplied in the same phase as the induced electromotive force, the current will have a phase lag with respect to the induced electromotive force due to the influence of the inductive load, producing reactive power and a second harmonic that does not contribute to the output. There is a problem in that the current greater than the wave component affects heat generation (efficiency).

The research conducted by the present inventors has shown that the efficiency or speed-torque characteristics of a brushless DC motor can be significantly improved by eliminating the delayed phase, that is, by improving the power factor.
Furthermore, regarding the removal of harmonic currents in the windings, the square wave AC voltage applied to the windings is pulse width modulated so that the voltage value during the supply ON time becomes close to a sine wave. The purpose is to improve speed-torque characteristics and heat generation (efficiency) by supplying at the right timing.

[Means to solve the problem]

In order to achieve this objective, the brushless DC motor driving method of the present invention uses a fundamental wave sine component that is approximately equivalent to the stator winding current and induced electromotive force of the brushless DC motor.
The product of each of the s component signals is performed for a period of N-3 magnetic poles, that is, 2π in electrical angle, and the phase difference between the winding current and the induced electromotive force is detected from the ratio, so that this phase difference is always approximately 0. Through closed-loop control, the ideal phase angle of the voltage applied to the winding with respect to the induced power is determined, the application timing is adjusted, and either the DC chopper method or the pulse width modulation method is controlled from time to time according to the applied voltage timing. The present invention is characterized by a driving method that provides a winding current that corresponds to an ideal phase angle and has few harmonic components by outputting a winding supply applied voltage selected according to the purpose.

[Effect]

The method for adjusting the switching timing of the voltage applied to a winding according to the present invention will be explained below.

FIG. 2(a) shows a vector diagram of the induced electromotive force E and the winding current ■ in the applied voltage method at normal timing. ℃), in order to eliminate the phase delay of the current i as shown in Figure 1(b), the voltage applied to the winding is advanced by λ6 and supplied, so that the phase difference between the induced electric power E and the winding current I is p in (a) - (b)? .

This shows a state in which the phase difference between the induced electric power E and the winding current (2) has not reached nearly zero, although it has become small. this? As a means of determining this, multiple reflectors or slits (both on the rotating side) as shown in Figure 4 (a) and a sensor boat (on the fixed side) as shown in (b) that detects this black and white level are combined in one pole. (For example, the higher the number, such as 60 or 120, the better the resolution and the smoother the control.) Place the magnet in the motor to obtain pulses in the N-3 section (3 in electrical angle).
60°) and a number of cosine signals having a phase difference of 90° from a sine signal corresponding to the previous pulse are obtained in advance and stored in a memory table so that they can be retrieved each time a pulse arrives.

These sin signals (e s), , cos signals (ec
) and the winding current i (including harmonic components) are expressed by the following equation.

e, -Σ E,%・sin nθ・
...(1) ec=ΣE, ・cosn(θ+T−>
・−・−・・(2) i=Σ I, −sin(m
θ+? ) ”...(3) Now, if we perform the signal processing of e, -i and ec-i (integration every 2π in electrical angle); -ΣE(' I K ' C05px (K: I+2
+”')4π k (4) However, if C3 does not include harmonic components, Ez=
Es=・・・・・・=0・・・・・・
(5) Similarly, it becomes possible to detect the phase difference 9 between the induced electromotive force and the winding current from the winding current i and es+ec containing harmonic components. This phase difference? The ideal deviation angle λ can be found by performing proportional + integral control so that λ is always approximately 0.

The switching timing of the U, V, and W phase supply voltages corresponding to this λ0 is determined, and the voltages are applied to the motor windings.

As a result, the phase of the induced power and the winding current are almost the same, the power factor is improved, and the speed-torque characteristics are significantly improved even with the same supply voltage.

Furthermore, when a pulse width modulated voltage is applied, the ratio of harmonic components flowing through the windings shifts to a higher order, resulting in a reduction in heat loss and contributing to an improvement in the efficiency of the entire motor.

In this way, the switching of the brushless DC motor can be shifted by the phase shift angle λ°, and the voltage applied to the winding can be changed to PW.
By converting to M, it becomes possible to significantly improve the speed-torque characteristics and efficiency of the motor.

〔Example〕

Hereinafter, the present invention will be specifically explained based on Examples. FIG. 5 shows an embodiment of the present invention.

In the figure, 12 is a brushless DC motor, 11 is a disc for detecting the position of a permanent magnet, and 10 is a group of reflective sensors for detecting the black and white level of this disc. This detection signal is input to the digital input section interface of the microcomputer board indicated by the dotted line box A. These detailed operations are shown in FIG. 4(a).

This will be explained in (b).

(a) is a reflective disc corresponding to that shown in FIG. 3, with the N pole and S pole of the permanent magnet arranged in the circumferential direction, with the N pole on the white side and the S pole on the black side. This method is based on signal level matching.

Figure 4(a) is an example of a 6-pole motor, and here the S pole ~
Although only the relationship between the N pole and the S pole is shown, there are actually 6 poles N- on the circumference.
They are formed in pairs such as 3-N-3, and each pole is divided into mechanical angles of 60 degrees. 2 in the diagram
(white) indicates the north pole and 3 (black) indicates the south pole. Also 1 is 6 for each pole
Divided into 0, every other part is white.

A light-reflective sensor whose surface is coated in black, white, black, etc., and a pulsed output is ON.
~OFF is installed so that 30 pulses are output in pairs. Therefore, the mechanical angle is 2 degrees per pulse, and 30 pulses from clock sensor 5 (CI) of 6 (C2
) outputs 30 pulses, a total of 60 pulses. The reason why two clock sensors are used is that when calculating the current phase retard angle, data is loaded every 3 degrees in electrical angle (1 degree in mechanical angle) from sin θ in Figure 5 and COS memory table.
This is a clock signal for the 1st N-pole section in Figure 4, and requires 60 pulses between the N-poles in Figure 4 (2), but due to the field of view and distance of the sensor, it is difficult to divide 120 pulses into 1 N-pole section due to the sensor specifications. Due to the sampling time due to the number of revolutions and clock number, that is, the processing speed of the microcomputer, here we divide it into 60 and divide the 30 pulses into the sensor 5 (
C1), the remaining 30 pulses are generated by sensor 6 (C2). However, sensor 5 and sensor 6 are 0 in mechanical angle.
．． They are installed with a phase difference of 5'', and addition processing is performed by the digital input section interface shown in FIG. 1, so that a clock of 60 pulses can be outputted per pole.

In addition, the discs 2 in Fig. 4(a) are attached corresponding to the positions in the circumferential direction of the permanent magnets of the 12 brushless DC motors.
(N pole) or 3 (S pole) sensor 8 (B), 7 (U), 4 (W)
are placed every 40 degrees in mechanical angle (120 degrees in electrical angle), and 3 light reflection sensors are placed in order to detect the N-3 section 8 (SW timing) of the permanent magnet, for a total of 5 light reflection sensors. . The situation is shown in FIGS. 3, 1O and 9, and FIG. 4(b).

CLKI and CLK2 are converted to 60 pulses/pole via the digital input interface l shown in Figure 5.
It is also input to the nθ and cosθ memory table 8 and the current phase retard detection unit 3, so that each value of sinθ and cosθ shown in FIGS. 6(a) and (b) can be retrieved in synchronization with the pulse in (C). Stored sinθ, COS
It is picked up from the θ memory table 8 and input to the current phase retard detection section 3. Also, brushless DC motor 1
The U-phase current of the winding current of No. 2 is detected by the CT type current sensor 13, and is manually input to the analog input section interface 2.
Furthermore, it is input to the current phase retard angle detection section 3.

Retardation of winding current with respect to induced power? In order to find
The U-phase on/off timing signal of the sensor 7 shown in FIG. 6(d) in FIG. (4) Calculate the product e of (i-th clock of sin θ) and winding current iu and the product of e (, (i-th clock of cos) and winding current iu. It is integrated over a 2π interval in terms of the electric angle between the poles.Here, it is assumed that e and e do not include harmonic components, so when the calculation is completed for the entire 2π interval, (6
) and (7), and the winding current is retarded with respect to the induced power? is determined from equation (8), and the result tan p is input to the subtracter 26. A target value is set to the other input of the subtracter. In other words, retard angle? =0 and tan p
= 0 is input as the target value. As a result, the deviation output is Δtan? ""tan p * -tan p, and the PI is adjusted so that this deviation Δtan p becomes approximately O.
It is input to the control system 5.

As mentioned above, Σe, ta and Σ in equations (4) and (5)
ec・iu is a 2π interval calculation for each clock, and at the end of the interval, the retardation tan of the winding current (p is the current phase retardation detection unit 3
It is output from Therefore, the deviation Δtan to the PI control section
The input of p is performed every 2π, and the timing of the voltage applied to the winding is calculated by the proportional + integral control system so that the induced power is advanced by λ° so that Δtan p = O, and the output control is performed. The signal is also input to section 6 every 2π. Output control section 6
is stored in the SW timing memory table 7 for DC chopper or PWM in the form of N=1°S=0 every 3° electrical angle in the state shown in Fig. 6(d) and (f), and the N pole -60・～N pole～S pole～S pole +60°, total 4
It stores 161 pieces of data for 806 minutes.

Before describing a method for controlling the winding voltage application timing using these data, the overall control strategy will be explained.

(1) The computation period required to perform one control output is the clock pulse period from the first rising of N pole (white) to the falling of S (black).

(2) The timing of the control output is the rising edge of the N pole (white).

(3) When the motor is stationary, the U-phase sensor and the permanent magnet position detection reflector do not necessarily coincide with the rise of 2 in FIG. 4(a). Therefore, advance angle control is not performed until the rotation occurs at the rising edge of the N pole, and U, V in Fig. 4(b),
W phase sensor? , 8.4 The signal that detected the magnet position is directly sent to the SW control unit 14 via the digital input unit I/Fl ~ initial setting unit 4 ~ digital output unit 1/F9, and the raw signal rotates the motor. let

(4) Clock sensor 5.6 of C1 and C2 in Figure 4 (b) has a resolution that changes depending on the light emission angle, viewing range, and distance of the optical sensor. Make the black mechanical angle narrower than necessary (this will make it impossible to detect it as a pulse, so when the diameter of the motor becomes small, install the C1 (5) sensor at a mechanical angle of 0.5 degrees electrical angle relative to the position). The digital input section 1/
In F1, we will create 60 equally divided pulses in the north pole and 60 equally divided pulses in the south pole, but if the diameter of the motor is large, one sensor will generate 60 pulses for each north and south pole from the reflective panel. shall be.

Next, actual advance angle control will be explained in detail.

A combination of a reflector attached to the permanent magnet position in Figure 5, a clock reflector 11 (details are shown in 1, 2, and 3 in Figure 4(a)), and a sensor mounting board in Figure 4(b). The permanent magnet positions and clock timings of the U, V, and W phases are detected and input to the digital input section 1/Fl in FIG. 5. As shown in (B) and (C) of Figure 6,
The U-phase winding current of the brushless DC motor shown in the figure is detected by a CT type sensor 13 and input to the analog input section 1/F2.

Clock CL detected from brushless DC motor 12
KI and 2 are processed into double period pulses by the digital input section 17F, and 30 pulses become 60 pulses in both the N-pole and S-pole sections. 3. Here, Σi u * sin θ and Σi, at each clock timing from the rising of the N pole of the U phase to the falling of the S pole,
* Cos θ is calculated and integrated by 2π in electrical angle, and the retard angle p is obtained in the form of tan p using equation (8). This p is the delay angle between the induced electromotive force e and the winding current i, and by phase control, history = 0, that is, tan p =
The proportional-integral (PI) control unit 5 advances the winding application timing by λ° so that the winding voltage is applied to the winding. This supply method is to calculate the deviation position (k) from the current point in the chopper/PWM SW timing memory table 7 corresponding to the advance angle λ° in the output control unit 6, and then input it to the current point X3 in the table 7. The data of the advanced point is loaded and output to the digital output section 1/F9, but since it is actually a 3-phase Y connection, the remaining ■ and W phases are ±2/2/2 with respect to the phase of the U phase.
Data at positions shifted by 3π are also picked up from the memory table 7 at the same time. Repeat this output every clock,
Repeat until the S pole (black) falls. After the end of the 2π interval, the retard angle p (tan p ) is calculated using equation (8), the above-mentioned items are repeatedly calculated, and a voltage is applied to the winding at the advanced timing.

By repeating these steps, when the control is started, the phase of the induced electric power e and the winding U-phase current 11 is delayed due to the inductive load, resulting in a low power factor, but as the rotation progresses, the delay angle also decreases, and eventually The deviation angle is controlled to be approximately O by a proportional-integral control system.

Therefore, the SW control unit 14 has λ for the U-phase induced power e.
The winding application timing pulse advanced by .degree. is also simultaneously inputted to the W phase, and the application timing pulse whose phase is shifted by ±2/3π is simultaneously inputted to the W phase. Also positive/with three timing pulses
A reverse signal is also input, and pace driver 15's U, V,
W and U, V.

One of W is selected and input to the power bridge section 16, and the upper (Upper) of the U, V, and W phases in FIG.
or Lower is selected and a voltage is applied to the motor coil at the timing when the angle is advanced, the motor rotates, the positions of the 51110th and 11th reflectors move, and the above-mentioned operation is repeated every 2π. become.

Furthermore, two methods of supplying DC voltage are being considered.

One is a PWM (pulse width modulation) method, and the other is a DC chopper method. These are the interlocking 5W18.1 in Figure 5.
9, and the selection signal from 5W19 is input to the output control unit 6 to select either the DC chipper or PWM SW timing memory table at 7, which corresponds to λ° during advance angle control. The SW pulse is supplied to the digital output unit I/F 9 to the SW control unit 14 with the advance angle position data in advance. An example of a pulse width train during PWM is shown in Figure 6 (
The signal (f) shown in (f) is an example of the timing chart (normal application timing, not advanced) of the upper side transistor TR+ of the U phase in FIG. 7, and the lower side transistor TR, . All of the N-pole sections in (d) are turned off. The upper and lower transistors TR and □T Rz t are handled in the same way as the U phase in a state where the phases are shifted by ±2/3π with respect to the phases of the V, W, and beam phases.

On the other hand, the DC chopper titrates the voltage (constant value) of the DC power supply 17 with the volume of the duty control 24 shown in the dotted line frame C, and converts the pulsed voltage into a constant voltage (filtering) with the coil and capacitor as shown in Fig. 6 (b). A voltage pulse like this will be applied. These PWM
, the DC chopper method is selected depending on the purpose.

As mentioned above, by applying the voltage to the winding without delay to the winding phase current 1 for the induced power e, the power factor cosφ
! =il, and the efficiency was greatly improved.

FIG. 8 shows an example of the speed-torque characteristics of a loKW class brushless DC motor in the DC chopper system, where the solid line indicates the advance angle control is present, and the broken line indicates the advance angle control heat. As is clear from the figure, the larger the voltage, the greater the effect; as the load increases, the required torque increases, and as a result, a large current is required, resulting in a larger reactive power. Due to the increase in

FIG. 9 shows a comparison of efficiency when measuring the speed-torque characteristics in FIG. 8, where the solid line represents the advance angle control and the dotted line represents the advance angle control heat. Efficiency is much higher with lead angle control.

As mentioned above, the phase difference between the U-phase current containing harmonic components and the in-phase induced electromotive force sine signal color cos signal signal and the upstream induced electromotive force and the winding current? Is this always detected? By configuring proportional+integral control so that ζ0 is achieved, the ideal advance angle λ of the winding applied voltage and induced electromotive force can be determined.

Therefore, the phase difference between the induced power and the winding current is approximately O.
The power factor became cos pζ1, and as shown in FIGS. 8 and 9, the speed-torque characteristics and the efficiency-torque characteristics were significantly improved.

Note that the ■ phase and the W phase are also carried out using the same concept as the U phase, so the details will be omitted.

〔Effect of the invention〕

As described above, in the present invention, by controlling the phase difference between the current flowing through the winding and the induced electromotive force of the winding to be approximately O at all times, the switching timing of the voltage applied to the winding can be adjusted. That is, the switching timing can be adjusted and both square wave and PWM methods can be used.

Therefore, as shown in Fig. 8, in the dotted line (no control) of the normal brushless DC motor drive method, as the voltage increases, the torque increases and the rotation speed decreases rapidly, and the speed-torque characteristic is similar to that of a general DC motor. However, by performing phase difference control, the speed-torque characteristic (solid line) becomes a nearly parallel straight line characteristic for each voltage, and speed control like a normal DC machine is possible. In other words, it means that maintaining the same torque and speed can be achieved with less voltage or current, which reduces heat loss and improves the overall efficiency of the motor.

Therefore, according to the present invention, the speed-torque characteristics and efficiency are greatly improved, and the effect is extremely large when used in applications requiring wide range of speed fluctuations and load fluctuations.

[Brief explanation of drawings]

Figure 1 (a) shows the voltage of a small capacity brushless DC motor,
Figure 2 (a) is a vector diagram of the application timing when there is no lead angle for the induced power e, (b) shows a vector diagram when the current phase lag remains to some extent but the angle is advanced. FIG. 3 is a sectional view of the brushless DC motor of this embodiment. Figure 4 (a) is an enlarged view of the clock generation reflector plate from Figure 3 and Figure 5, (b) is an enlarged view of the permanent magnet position detection board, Figure 5 is a block diagram for advance angle control, Figure 6 (a) and (b) are sin θ and cos in FIG.
θ A diagram that continuously represents the state of the memory table on the time axis.
) is the permanent magnet position detection timing, (e) is the applied voltage timing when advanced by λ°, and (f) is the S during PWM.
Figure 7 is a circuit diagram showing the W timing, Figure 7 is a circuit diagram showing the Y-connected windings and the power bridge section for driving them, Figure 8 is a comparison of speed-torque characteristics with and without advance angle control. The graph in FIG. 9 is an explanatory diagram showing a comparison of efficiency and torque characteristics during the test in FIG. 8. Applicant Nippon Steel Corporation Representative Patent Attorney Minoru Aoyagi 4 5
□ Figure 3 Rotation I: (-to day) Meng Gofan Request for amendment (method) 1. Indication of the case 1999 Patent Application No. 120014 Address 2-6-3 Otemachi, Chiyoda-ku, Tokyo Name (6
65) Nippon Steel Corporation Representative Yutaka Saito 4, Agent 101 fi' 0
3 (863) 02206, Number of claims increased by amendment None 7, Subject of amendment Drawings

Claims (1)

[Claims] 1. An encoder (Nφ
1 is for N pole and S pole detection, Nφ2 is a metal plate for detecting current phase delay calculation timing clock), and black and white labels are detected by optical sensors for magnetic pole position and timing clock for each U, V, W phase. Among the total number of poles, a pair of N and S poles (2π in electrical angle) is configured so that the required number of clocks can be generated in this section, and the in-phase induced power calculated in advance for each clock is The sine value and cosine value are taken out from the table, each is multiplied by the U-phase current containing high frequency components, and integrated. When the calculation for one section is completed, the current phase delay angle is determined, and this is calculated every 2π period. To ensure that the phase lag angle at that time is always almost 0 using a proportional + integral control system, we have prepared a switching memory table (for pulse width modulation (PWM) and DC U phase data at a position where the phase is shifted by the number of pieces converted to the minimum resolution of the clock (V and W phases are further ±2/3π with respect to the U phase). By extracting the misaligned data and outputting it to the switching control unit, the delay of the fundamental wave current of each phase with respect to the phase of the induced power is reduced, the power factor becomes almost 1, and the efficiency of the entire motor is greatly improved. A method for driving a brushless DC motor, characterized by: