Semiconductor device, motor drive system, and method of starting a motor

Abstract

The present application relates to a semiconductor device, a motor drive system, and a method of starting a motor. A rotor magnetic pole position is estimated with high accuracy at the time of initial start of a three-phase motor of a sensorless system. A semiconductor device for driving and controlling a three-phase motor of a sensorless system has a detector connected to three-phase output nodes and a virtual neutral point (or neutral point) of an inverter circuit, and detects a voltage generated in the output node of a non-energized phase among the three phases. A controller applies an initial drive voltage to any two phases of the three-phase motor through the inverter circuit based on the estimated magnetic pole position of the rotor in a stopped state. The controller estimates the position of the rotor based on a differential voltage detected by the detector in a drive voltage application period and in a regeneration period immediately after or before the drive voltage application period.

Description

[0001] Cross-references to related applications

[0002] This document incorporates by reference the full disclosure of Japanese Patent Application No. 2020-165185, filed on September 30, 2020, including the specification, drawings, and abstract. Technical Field

[0003] This invention relates to semiconductor devices, motor drive systems, and methods for starting motors. For example, it is applicable to the initial drive of brushless DC motors (also known as permanent magnet synchronous motors) with sensorless systems. Background Technology

[0004] In sensorless control of brushless DC motors, the back electromotive force (BEMF) generated in the stator windings of the non-energized phase is detected by the rotation of the motor rotor. Based on the zero-crossing point of the detected BEMF, the rotor position and speed are estimated.

[0005] Since no back electromotive force is generated when the motor stops, the state of the motor rotor cannot be estimated using the sensorless control described above, which is based on detecting the zero-crossing point of the back electromotive force. Therefore, for example, the rotor's magnetic pole position is estimated by using the self-inductance difference.

[0006] Subsequently, the motor is initially driven by applying a drive voltage to the stator winding of the appropriate phase corresponding to the estimated initial magnetic pole position for a predetermined time. Multiple initial drives are performed while switching the phase used to apply the drive voltage. When a sufficient amount of back electromotive force is detected, motor control switches to sensorless control based on the zero-crossing point of the detected back electromotive force.

[0007] The publicly available technologies are listed below.

[0008] [Patent Document 1] Japanese Unexamined Patent Application Publication No. 2008-113506.

[0009] Patent Document 1 discloses a method for estimating rotor position during initial drive to shorten motor start-up time. Specifically, the motor drive control device of Patent Document 1 detects the peak value of the induced voltage generated in the stator winding of the non-energized phase due to mutual inductance with the stator winding of the energized phase. The motor drive control device then switches the energized phase based on comparing the detected peak value of the induced voltage with a threshold value. Summary of the Invention

[0010] In the motor starting method disclosed in Patent Document 1, it is difficult to set a threshold for comparison with the peak value of the induced voltage. This is because the voltage generated in the non-energized phase includes not only the induced voltage caused by mutual inductance but also a small amount of induced voltage based on back electromotive force. Furthermore, due to individual differences in motors, the amplitude of the induced voltage generated by mutual inductance is relatively large. Therefore, it is difficult to accurately detect the rotor position.

[0011] Other objects and novel features will become apparent from the description in this specification and the accompanying drawings.

[0012] According to an embodiment of the semiconductor device, during the initial startup of the motor, the semiconductor device detects the difference between the voltage generated in the non-energized phase during a drive voltage application period and the voltage generated in the non-energized phase during a regeneration period immediately following or preceding the drive voltage application period. The semiconductor device estimates the rotor's magnetic pole position based on the detected voltage difference.

[0013] According to the above embodiments, the position of the rotor's magnetic poles can be estimated with high accuracy during the initial start-up of the motor, without being affected by back electromotive force and individual differences in the motor. Attached Figure Description

[0014] Figure 1 This is a block diagram illustrating an example configuration of a motor drive system.

[0015] Figure 2 It is shown Figure 1 A block diagram illustrating an example configuration of an MCU.

[0016] Figure 3 This diagram illustrates six energization modes under the 120-degree energization method.

[0017] Figure 4 This is a timing diagram showing the waveforms of the gate control signals provided to each MOS transistor.

[0018] Figure 5 This is a flowchart illustrating an example of the driving process of a brushless DC motor.

[0019] Figure 6 yes Figure 1 Examples of current waveforms for each phase of a motor drive system.

[0020] Figure 7 It is shown Figure 5 The flowchart shows the detailed operation of the mutual inductance detection drive in step S60.

[0021] Figure 8A and Figure 8B yes Figure 1 An example of the output waveform of a differential amplifier in a mutual inductance detection drive.

[0022] Figure 9A , Figure 9B and Figure 9C It is used to explain in Figure 8A and Figure 8B A diagram showing the rotor positions at points α, β, and γ. Detailed Implementation

[0023] In the following description, embodiments will be described in detail with reference to the accompanying drawings. Identical or corresponding parts are indicated by the same reference numerals and will not be described again thereto.

[0024] Overall configuration of the motor drive system

[0025] Figure 1 This is a block diagram illustrating an example configuration of a motor drive system. (Refer to...) Figure 1 The motor drive system 100 includes a three-phase brushless DC motor 30, an inverter circuit 20, and a semiconductor device 10 for controlling the inverter circuit 20.

[0026] 1. Brushless DC motor

[0027] The brushless DC motor 30 includes Y-connected stator windings 31U, 31V, and 31W and a rotor having one or more pole pairs (not shown). The rotor is driven to rotate synchronously with the three-phase AC power supplied from the inverter circuit 20 to the stator windings 31U, 31V, and 31W. The node of the stator windings 31U, 31V, and 31W is called the neutral point 32.

[0028] In the following text, Figure 1 The case where the stator winding 31 is Y-connected is shown, but the technology of this disclosure has the same effect when the stator winding 31 is Δ-connected.

[0029] 2. Inverter circuit

[0030] The inverter circuit 20 includes MOS (metal-oxide-semiconductor) transistors UP, UN, VP, VN, WP, and WN.

[0031] MOS transistor UP is located on the upper arm of phase U, and MOS transistor UN is located on the lower arm of phase U. MOS transistor VP is located on the upper arm of phase V, and MOS transistor VN is located on the lower arm of phase V. MOS transistor WP is located on the upper arm of phase W, and MOS transistor WN is located on the lower arm of phase W. The upper arm is also called the high side, and the lower arm is also called the low side.

[0032] In addition, the inverter circuit 20 includes a shunt resistor 23 for detecting the current between the two phases.

[0033] The following is a brief description of these connections. MOS transistors UP and UN are connected in series between the first power supply node 21 and the connection node 24, arranged in this order. The first power supply node 21 is supplied with an external input voltage VM (also referred to as the power supply voltage VM), and node 24 is connected on the low-potential side. Shunt resistor 23 is connected between connection node 24 and the second power supply node 22, which provides the ground voltage GND. Output node TU is the connection node between MOS transistors UP and UN. Output node TU is connected to one end of the U-phase stator winding 31U.

[0034] Similarly, MOS transistors VP and VN are connected in series between the first power supply node 21 and the connection node 24 in this order. The output node TV is the connection node between MOS transistors VP and VN. The output node TV is connected to one end of the V-phase stator winding 31V.

[0035] Similarly, MOS transistors WP and WN are connected in series between the first power supply node 21 and the connection node 24 in this order. The output node TW is the connection node between MOS transistors WP and WN. The output node TW is connected to one end of the W-phase stator winding 31W.

[0036] Each MOS transistor UP, UN, VP, VN, WP, WN has a reverse-biased body diode (not shown) connected in parallel. Therefore, when both the upper and lower arm transistors on the same phase are in the off state, regenerative current flows through the body diode.

[0037] exist Figure 1 In this configuration, all MOS transistors UP, UN, VP, VN, WP, and WN are composed of N-channel MOS transistors. Alternatively, any one of the MOS transistors UP, VP, WP, UN, VN, and WN in the upper arm can be an NMOS. In this case, the other MOS transistors are PMOS. Alternatively, all MOS transistors UP, UN, VP, VN, WP, and WN can be composed of P-channel MOS transistors.

[0038] In addition to MOS transistors, other types of field-effect transistors can be used as semiconductor switching elements constituting inverter circuit 20. Alternatively, bipolar transistors or insulated-gate bipolar transistors (IGBTs) can be used instead of MOS transistors. However, when using other types of transistors, the flywheel diodes need to be connected in anti-parallel with each transistor. This is done to allow current to flow through the regeneration path when both the upper and lower arms of the transistors on the same phase are in the off state.

[0039] 3. Semiconductor devices

[0040] The semiconductor device 10 includes a switching circuit 61, a virtual neutral point generation circuit 70, a differential amplifier 63, an amplifier 65, and a microcontroller unit (MCU) 40.

[0041] The switching circuit 61 and the differential amplifier 63 constitute a detector 60, which is used to detect the voltage of the output node of the non-energized phase of the inverter circuit 20.

[0042] Switching circuit 61 is connected to output nodes TU, TV, and TW. In response to the phase selection signals SLU, SLV, and SLW output from MCU 40, switching circuit 61 connects detection node 62 to one of the output nodes TU, TV, and TW corresponding to the selected phase.

[0043] The virtual neutral point generation circuit 70 generates a virtual neutral point 72 having the same voltage as the neutral point 32 of the brushless DC motor 30. The virtual neutral point generation circuit 70 includes resistive elements 71U, 71V, and 71W. Resistor 71U is connected between the virtual neutral point 72 and the output node TU. Resistor 71V is connected between the virtual neutral point 72 and the output node TV. Resistor 71W is connected between the virtual neutral point 72 and the output node TW. Resistors 71U, 71V, and 71W have equal resistance values.

[0044] Differential amplifier 63 amplifies the difference between the voltage Vd at detection node 62 and the reference voltage Vref. The voltage at neutral point 32 or virtual neutral point 72 is used as the reference voltage Vref.

[0045] Amplifier 65 amplifies the voltage appearing in shunt resistor 23. Therefore, the motor current flowing between phases U, V, and W can be detected.

[0046] The MCU 40 integrates a computer, including a CPU (Central Processing Unit) and memory, into a single integrated circuit. The MCU 40 performs various functions by executing programs stored in its memory. References will now be made. Figure 1 and Figure 2 This describes the configuration and functions of MCU 10.

[0047] Figure 2 It is shown Figure 1 A block diagram illustrating an example configuration of the MCU. Figure 2 In this configuration, the MCU 40 is configured by a computer including a CPU (Central Processing Unit) 41, RAM (Random Access Memory) 4, and non-volatile memory 43.

[0048] The MCU 40 also includes interface (IF) circuits 44 and 46, analog-to-digital (AD) converters 45 and 48, and gate control signal generation circuitry 47. The MCU 40 also includes a bus 49 for interconnecting these components. Two or more components, such as a CPU 41, may be provided.

[0049] CPU 41 operates according to instructions contained in a program stored in non-volatile memory 43, thereby controlling the entire motor drive system 100. This program may be provided as a non-transient storage medium or via a network.

[0050] RAM 42 serves as the main memory for CPU 41 operations. Non-volatile memory 43 stores the program and various settings required for program operation. Non-volatile memory 43 can be a mask ROM (random access memory) or EEPROM (electrically erasable programmable ROM), etc. Alternatively, non-volatile memory 43 can be NOR flash memory, NAND flash memory, etc. Non-volatile memory 43 may also include SSD (solid-state drive) or hard disk.

[0051] IF circuit 44 is used to receive the input of operation command value 11 from the outside. IF circuit 46 is used to output phase selection signals SLU, SLV, and SLW to switching circuit 61 based on commands from CPU 41. For example, IF circuits 44 and 46 provide input / output isolation, leveling, and timing adjustment between the internal and external circuits of MCU 40.

[0052] AD converter 45 converts the output signal Vout of differential amplifier 63 into a digital signal. AD converter 48 converts the output signal Id of amplifier 65 into a digital signal. AD converters 45 and 48 can have any known circuit configuration.

[0053] The gate control signal generation circuit 47 generates gate control signals GUP, GUN, GVP, GVN, GWP, and GWN for the MOS transistors UP, UN, VP, VN, WP, and WN based on commands from the CPU 41. For example, the gate control signal generation circuit 47 generates gate control signals based on PWM (Pulse Width Modulation) control by comparing the gate voltage command value of each MOS transistor with the value of the carrier signal.

[0054] Instead of the MCU 40 with the above configuration, a controller configured with dedicated circuitry (such as an ASIC (Application-Specific Integrated Circuit)) can be used. Alternatively, at least two of ASICs, FPGAs (Field-Programmable Gate Arrays), CPUs, etc., can be combined to configure the controller.

[0055] Power-on mode

[0056] Next, six power-on modes will be described using the 120-degree power-on method. In this embodiment, the MCU 40 controls the brushless DC motor 30 using the 120-degree power-on method. The 120-degree power-on method uses a 120-degree electrical angle half-cycle as the power-on period and the remaining 60-degree electrical angle half-cycle as the power-off period. BEMF can be detected during the power-off period. In a three-phase brushless DC motor, since the power-on phase switches every 60 electrical angles, there are six power-on modes.

[0057] Note that the disclosed technique can be applied as long as the back electromotive force generated during the non-energized period can be measured, even when the energized period is greater than 120 degrees and less than 180 degrees of the electrical angle half-cycle.

[0058] It should be noted that the power-on and power-off periods of the above-mentioned 120-degree power-on method are different from the voltage application period (also known as the turn-on period) and regeneration period (also known as the turn-off period) of PWM control. The power-on period includes multiple voltage application periods and multiple regeneration periods of PWM control.

[0059] Figure 3 This diagram illustrates six energizing modes under the 120-degree energizing method. Figure 3 The diagram shows six energizing modes of current af.

[0060] (a) Reference Figure 3 In this configuration, the MOS transistor WP in the upper arm of phase W and the MOS transistor VN in the lower arm of phase V are turned on, while the other transistors are turned off. Therefore, the motor current flows from the stator winding 31W in phase W to the stator winding 31V in phase V. When the stator winding 31U in phase U is de-energized, the zero-crossing point of BEMF can be observed. In the following description, this energizing mode is referred to as energizing mode a.

[0061] Furthermore, in energizing mode a, phase U is referred to as the "non-energized phase," phase W as the "upstream energized phase," and phase V as the "downstream energized phase." The motor current flows in the direction from the stator windings of the upstream energized phase to the stator windings of the downstream energized phase. The same definition applies to other energizing modes.

[0062] When PWM control is performed in power-on mode a, MCU 40 keeps the upper arm of phase W on and performs PWM control on phase V. Alternatively, MCU 40 keeps the lower arm of phase V on and performs PWM control on phase W. In the former case, regenerative current flows through the upper arm of phase W and the upper arm of phase V. In the latter case, regenerative current flows through the lower arm of phase W and the lower arm of phase V.

[0063] (b) In inverter circuit 20, the MOS transistor WP in the upper arm of phase W and the MOS transistor UN in the lower arm of phase U are controlled to be turned on, while the other transistors are controlled to be turned off. Therefore, the motor current b flows from the stator winding 31W of phase W to the stator winding 31U of phase U. The stator winding 31V of phase V is in a non-energized state, and the zero-crossing point of BEMF can be observed. In the following description, this energizing mode is referred to as energizing mode b.

[0064] When PWM control is executed in power-on mode b, MCU 40 keeps the upper arm of phase W on and performs PWM control on phase U. Alternatively, MCU 40 keeps the lower arm of phase U on and performs PWM control on phase W. In the former case, regenerative current flows through both the upper arm of phase W and the upper arm of phase U. In the latter case, regenerative current flows through both the lower arm of phase W and the lower arm of phase U.

[0065] (c) When the MOS transistor VP of the upper arm of phase V and the MOS transistor UN of the lower arm of phase U are turned on and the other transistors are turned off, the motor current c flows from the V-phase stator winding 31V to the U-phase stator winding 31U. The W-phase stator winding 31W is in a non-energized state, and the zero-crossing point of BEMF can be observed. In the following description, this is referred to as energizing mode c.

[0066] When PWM control is performed in power-on mode c, MCU 40 keeps the upper arm of phase V on and performs PWM control on phase U. Alternatively, MCU 40 keeps the lower arm of phase U on and performs PWM control on phase V. In the former case, regenerative current flows through both the upper arm of phase V and the upper arm of phase U. In the latter case, regenerative current flows through both the lower arm of phase V and the lower arm of phase U.

[0067] (d) When the MOS transistor VP of the upper arm of phase V and the MOS transistor WN of the lower arm of phase W are turned on and the other transistors are turned off, the motor current d flows from the V-phase stator winding 31V to the W-phase stator winding 31W. The U-phase stator winding 31U is de-energized, and the zero-crossing point of BEMF can be observed. In the following description, this energizing mode is referred to as energizing mode d.

[0068] When PWM control is performed in power-on mode d, MCU 40 keeps the upper arm of phase V on and performs PWM control on phase W. Alternatively, MCU 40 keeps the lower arm of phase W on and performs PWM control on phase V. In the former case, regenerative current flows through both the upper arm of phase V and the upper arm of phase W. In the latter case, regenerative current flows through both the lower arm of phase V and the lower arm of phase W.

[0069] (e) When the MOS transistor UP in the upper arm of phase U and the MOS transistor WN in the lower arm of phase W are turned on and the other transistors are turned off, the motor current e flows from the stator winding 31U in phase U to the stator winding 31W in phase W. The stator winding 31V in phase V is in a non-energized state, and the zero-crossing point of BEMF can be observed. In the following description, this energizing mode is referred to as energizing mode e.

[0070] When PWM control is performed in power-on mode e, MCU 40 keeps the upper arm of phase U on and performs PWM control on phase W. Alternatively, MCU 40 keeps the lower arm of phase W on and performs PWM control on phase U. In the former case, regenerative current flows through both the upper arm of phase U and the upper arm of phase W. In the latter case, regenerative current flows through both the lower arm of phase U and the lower arm of phase W.

[0071] (f) When the MOS transistor UP in the upper arm of phase U and the MOS transistor VN in the lower arm of phase V are turned on and the other transistors are turned off, the motor current f flows from the stator winding 31U in phase U to the stator winding 31V in phase V. The stator winding 31W in phase W is in a non-energized state, and the zero-crossing point of BEMF can be observed. In the following description, this energizing mode is referred to as energizing mode f.

[0072] When PWM control is performed in power-on mode f, MCU 40 keeps the upper arm of phase U on and performs PWM control on phase V. Alternatively, MCU 40 keeps the lower arm of phase V on and performs PWM control on phase U. In the former case, regenerative current flows through both the upper arm of phase U and the upper arm of phase V. In the latter case, regenerative current flows through both the lower arm of phase U and the lower arm of phase V.

[0073] By controlling the inverter circuit 20 to make the current flow to the brushless DC motor 30 in the order of the above-described energizing modes a, b, c, d, e, f, the energizing phases are switched sequentially, and the rotor of the brushless DC motor 30 also rotates synchronously with the rotating electromagnetic field. In this specification, for convenience, this rotation direction is referred to as clockwise (CW).

[0074] On the other hand, when the power-on modes are switched in the reverse order, the power-on phases are also switched in the reverse order. That is, when the inverter circuit 20 is controlled to make the current flow to the brushless DC motor 30 in the order of power-on modes f, e, d, c, b, a, the power-on phases are switched in the reverse order. Therefore, the rotor of the brushless DC motor 30 also rotates synchronously with the rotating electromagnetic field. This direction of rotation is referred to as the counterclockwise rotation (CCW) direction in this specification.

[0075] Figure 4 This is a timing diagram showing the waveforms of the gate control signals supplied to each MOS transistor. Figure 4 The image shows an example of the waveform for each energizing mode when the regenerative current flows through the lower arm but not through the upper arm. Figure 4 The waveforms shown are conceptual and not entirely the same as actual waveforms. When the gate voltage is high (H), the corresponding MOS transistor is turned on. When the gate voltage is low (L), the corresponding MOS transistor is turned off. Figure 4 In this context, the non-energized phase is referred to as the high-impedance (high-Z) phase.

[0076] Overview of motor drive operation

[0077] In this embodiment, the motor drive operation is generally divided into three modes. Their outlines will be described below first.

[0078] In the first mode, MCU 40 estimates the initial position of the magnetic poles of the rotor in a stopped state. For example, inductive sensing is used to detect the initial magnetic pole position. In inductive sensing, for example, when... Figure 3 In the six energizing modes shown, the difference in current flowing through the stator winding 31 is detected when a voltage is applied to the stator winding 31 at the degree to which the rotor does not rotate. The current flowing through the stator winding 31 is detected by the voltage generated in the shunt resistor 23.

[0079] In the case of a brushless DC motor 30 with salient poles, the magnetic pole position can be detected based on the change in current in the windings because the inductance in the d-axis direction is reduced. Furthermore, in the case of a non-salient brushless DC motor 30, the magnetic pole position can be detected by detecting the decrease in inductance due to magnetic saturation based on the change in current.

[0080] Specifically, the current I flowing through the stator winding 31 is expressed by formula (1):

[0081] I=V / R[1-exp(-t·R / L)] (1)

[0082] Where R represents the winding resistance of stator winding 31, L represents the inductance, and V represents the applied voltage.

[0083] In the second mode, an initial starting torque is provided to the rotor via PWM drive based on the initial magnetic pole position detected in the first mode. Applying the starting torque is also called a kick. In this embodiment, to estimate the rotor's magnetic pole position, the mutual inductance voltage in the non-energized phase stator winding 31 generated by the current magnetic field of the energized phase stator winding 31 is detected. As described later, the mutual inductance voltage is obtained by PWM control of the difference between the non-energized phase induced voltage during the voltage application period and the non-energized phase induced voltage immediately after (or before) the voltage application period during the regeneration period. In this invention, the second mode is referred to as mutual inductance detection drive.

[0084] Specifically, the magnetic flux φ1 is represented by formula (2).

[0085] φ1=B1S=μn1I1S (2)

[0086] Where n1 represents the number of stator windings 31 in the energized phase, I1 represents the winding current, S represents the cross-sectional area of ​​the coil, μ represents the permeability, and B1 represents the magnetic flux density of the current magnetic field.

[0087] In addition, the magnetic flux φ2 of the non-energized phase is represented by equation (3), and the induced voltage e2 is represented by equation (4).

[0088] φ2∝φ1 / K (3)

[0089] e²∝n²·Δφ² / Δt (4)

[0090] Where n2 represents the number of turns of the stator winding 31 in the non-energized phase. K is the proportionality coefficient that varies with the rotor position, and Δφ2 / Δt is the time rate of change (i.e., derivative) of the magnetic flux φ2.

[0091] Figure 4 An example of the voltage waveforms of the gate control signals GUP, GUN, GVP, GVN, GWP, and GWN in the second mode is shown. That is, in the second mode, to improve the detection accuracy of the induced voltage during the regeneration period, it is desirable to control the inverter circuit 20 so that the regeneration current... Figure 4 The current flows through the lower arm as shown. However, if increased detection accuracy is not required, the regenerative current does not necessarily have to flow only through the lower arm. For example, the regenerative current can flow through the upper arm.

[0092] In the third mode, the brushless DC motor 30 is driven based on the magnetic pole position estimated according to the zero-crossing point of the BEMF. When the amplitude of the BEMF detected in the second mode reaches a predetermined amplitude, the driving mode changes from the second mode to the third mode. In this invention, the third mode is referred to as the BEMF detection driving mode.

[0093] Here, the induced voltage generated by BEMF refers to the voltage generated in the stator winding 31 due to the rotation of the rotor. Specifically, the induced voltage V generated by BEMF is expressed by formula (5):

[0094] V = Blv = 2πrNBl (5)

[0095] Where B represents the magnetic flux of the permanent magnet provided in the rotor, v represents the relative velocity between the rotor surface and the stator winding 31, r represents the rotor's radius of rotation, and N represents the rotor's rotational speed. For simplicity, assume the coil is a square with side l.

[0096] The waveforms of the gate control signals GUP, GUN, GVP, GVN, GWP, and GWN in the third mode can be used... Figure 4 The voltage waveform or other voltage waveform. For example, it can be modified. Figure 4 The voltage waveform causes the phase used to perform PWM control to be discontinuous.

[0097] Described Figure 4 Several modified examples of the waveforms are shown. In power-on mode b, phase U can be controlled by PWM, the gate control signal GWP can be controlled to level H, and the gate control signal GWN can be controlled to level L. In power-on mode d, phase W is controlled by PWM, and the gate control signal GVP is controlled to level H. Figure 4 The waveform is altered, causing the gate control signal GVN to be controlled to the L level. In power-on mode f, phase V is controlled by PWM, and the gate control signal GUP is controlled to the H level. Figure 4 The waveform is changed, causing the gate control signal GUN to be controlled to the L level. In this case, the waveforms of power-on modes a, c, and e remain unchanged.

[0098] Details of motor drive operation

[0099] Figure 5 This is a flowchart illustrating an example of the drive process for a brushless DC motor. See below for reference. Figure 1 and Figure 5 The drive process for the brushless DC motor 30 is explained below. The detected value of the external input voltage VM is input to the MCU 40.

[0100] exist Figure 5 In step S10, MCU 40 determines whether the external input voltage VM is equal to or higher than the preset startup voltage. When the external input voltage VM is equal to or higher than the startup voltage, MCU 40 causes the process to proceed to step S20.

[0101] In step S20, the MCU 40 sets the gate control signals GUP, GUN, GVP, GVN, GWP, and GWN to a low level. This causes the MOS transistors UP, UN, VP, VN, WP, and WN to be in a turned-off state (high impedance state).

[0102] In the next step S30, the MCU 40 detects the rotor's rotation direction by detecting the induced voltage (BEMF) of each phase while sequentially switching on the three phases of the switching circuit 61. In the MCU 40, when the rotor's rotation direction is forward according to the direction specified by the operation command value 11, the process proceeds to step S70 (BEMF detection drive). In the MCU 40, when the rotor's rotation direction is reversed, opposite to the direction specified by the operation command value 11, the process proceeds to step S100 (three-phase short braking). In the MCU 40, when the BEMF amplitude is equal to or less than a threshold, it is determined that the rotor is not rotating, and the process proceeds to step S40.

[0103] In step S40, the MCU 40 detects the initial magnetic pole position by, for example, using inductive sensing.

[0104] In the next step S50, the MCU 40 applies voltage to the stator winding 31 in an energized mode based on the initial magnetic pole position detected in step S40, where the initial torque increases to its maximum. For example, in step S40, it is assumed that minimum self-inductance is detected in energized mode d. In this case, in order to rotate along the CW direction, the MCU 40 controls the inverter circuit 20 to apply voltage to the stator winding 31 in energized mode f.

[0105] In step S60, MCU 40 executes the mutual inductance detection drive (mode 2). See below for further details. Figure 7 Figure 9 describes the details of the mutual inductance detection drive.

[0106] In step S60, when MCU 40 determines that the rotational speed is insufficient, the process returns to step S50, and the initial torque is applied in the next power-on mode. If rotor rotation cannot be detected even if the shift time exceeds the maximum shift time Tmax, MCU 40 determines a timeout. In this case, MCU 40 detects the initial magnetic pole position of the rotor by returning the process to step S40. When it is detected that the rotor speed has reached the specified speed, MCU 40 proceeds the process to step S70.

[0107] In step S70, MCU 40 executes BEMF detection drive (mode 3). Specifically, based on the zero-crossing point of the BEMF detected in the non-energized phase, MCU 40 estimates the rotor position and speed. MCU 40 drives the brushless DC motor 30 via PWM control to apply appropriate torque to the rotor based on the estimated rotor position and speed.

[0108] If no BEMF detection error is detected (No in step S80), and the external input voltage VM is equal to or higher than the startup voltage (No in step S90), MCU 40 continues to execute step S70. If a BEMF detection error is detected (Yes in step S80), and the external input voltage VM is less than the startup voltage (Yes in step S90), MCU 40 causes the process to proceed to step S100.

[0109] In step S100, the MCU 40 disconnects the external input voltage VM, preventing it from being supplied to the first power node 21. Then, the MCU 40 sets the gate control signals GUP, GUN, GVP, GVN, GWP, and GWN to H level, turning on all MOS transistors UP, UN, VP, VN, WP, and WN. Therefore, a three-phase short-circuit brake is applied to the brushless DC motor 30.

[0110] In step S110, the MCU 40 sets the phase selection signals SLU, SLV, and SLW, causing any phase of the switching circuit 61 to be turned on. In this state, for example, when the MCU 40 controls the U phase of the switching circuit 61 to be turned on, the MCU 40 monitors the terminal voltage between the output node TU and the virtual neutral point 72. The MCU 40 continues to perform three-phase short braking (step S100) until the amplitude (or maximum value) of the terminal voltage is equal to or less than a preset stop threshold. When the amplitude (or maximum value) of the terminal voltage becomes equal to or less than the stop threshold (yes in step S110), the MCU 40 returns the process to the first step S10.

[0111] Figure 6 yes Figure 1 Examples of current waveforms for each phase of a motor drive system. For ease of understanding, the amplitude and duration of the applied current may be distorted and not proportional to the actual current.

[0112] Reference Figure 6 The time interval from t0 to t10 corresponds to mode 1 (initial magnetic pole position detection). The time interval from t10 to t20 corresponds to mode 2 (mutual inductance detection drive). The time interval from t20 onwards corresponds to mode 3 (BEMF detection drive).

[0113] From time t0 to time t10, MCU 40 controls inverter circuit 20 according to... Figure 3The six energizing modes a to f described in the diagram are applied sequentially to the stator winding 31. In this case, the voltage applied to the stator winding 31 and its application time are limited to the extent that the rotor does not rotate. The MCU 40 detects the current flowing through the stator winding 31 in the corresponding energizing mode (i.e., the voltage generated in the shunt resistor 23). Based on the detection results, the MCU 40 determines the energizing mode when the self-inductance is lowest.

[0114] From time t10 to time t20, MCU 40 applies starting torque to brushless DC motor 30 by controlling the inverter circuit.

[0115] First, at time t10, MCU 40 applies a PWM-controlled voltage to the stator winding 31 in energized mode, with the maximum applied torque increasing based on the rotor position detected by inductive sensing. Figure 6 In the case of PWM control, firstly, a voltage is applied in energized mode f. During the voltage application period controlled by PWM, current flows through stator winding 31 from the upper arm of phase U to the lower arm of phase V. During the regeneration period controlled by PWM, regenerative current flows from the lower arm of phase U through stator winding 31 in the direction of the lower arm of phase V. MCU 40 detects the induced voltage of phase W (non-energized phase) via detector 60. MCU 40 estimates the mutual inductance voltage and BEMF based on the detected induced voltage of the non-energized phase. The following will refer to... Figure 7 Figure 9 discusses the specific methods used to estimate mutual inductance voltage and VEMF.

[0116] At time t11, the mutual inductance voltage detected in the non-energized phase W becomes essentially zero (i.e., the absolute value of the mutual inductance voltage is less than the threshold). At this time, since the absolute value of BEMF detected through the non-energized phase W is not equal to or greater than the threshold voltage Vth, MCU 40 switches the energizing mode. That is, MCU 40, by controlling inverter circuit 20, applies a PWM-controlled voltage to stator winding 31 in the next energizing mode a. In energizing mode a, during the PWM-controlled voltage application period, current flows through stator winding 31 from the upper arm of phase W to the lower arm of phase V. During the PWM-controlled regeneration period, regenerative current flows from the lower arm of phase W through stator winding 31 in the direction of the lower arm of phase V. MCU 40 detects the induced voltage of phase U, which is the non-energized phase, via detector 60.

[0117] At time t12, the mutual inductance voltage detected in the non-energized phase U becomes essentially zero (i.e., the absolute value of the mutual inductance voltage is less than the threshold). At this time, since the absolute value of BEMF detected by the non-energized phase U is not equal to or greater than the threshold voltage Vth, MCU 40 switches the energizing mode. That is, MCU 40 applies a PWM-controlled voltage to the stator winding 31 in the next energizing mode b by controlling the inverter circuit 20. In energizing mode b, during the PWM-controlled voltage application period, current flows through the stator winding 31 from the upper arm of phase W to the lower arm of phase U. During the PWM-controlled regeneration period, regenerative current flows from the lower arm of phase W along the direction of the lower arm of phase U through the stator winding 31. MCU 40 detects the induced voltage of phase V, which is the non-energized phase, via detector 60.

[0118] At time t13, the mutual inductance voltage detected in the non-energized phase V becomes essentially zero (i.e., the absolute value of the mutual inductance voltage is less than the threshold). At this time, since the absolute value of BEMF detected through the non-energized phase V is equal to or greater than the threshold voltage Vth, MCU 40 switches the drive mode from mode 2 to mode 3 at the next time t20.

[0119] At time t20, MCU 40 applies a PWM-controlled voltage to the stator winding 31 in power-on mode d by controlling inverter circuit 20. During power-on mode d, MCU 40 detects the zero-crossing point of the BEMF of the non-powered U phase via detector 60. BEMF in mode 3 is detected via CR filter.

[0120] At time t21, after 30 electrical degrees have elapsed since the zero-crossing detection time, MCU 40 switches the power-on mode from power-on mode d to power-on mode e. Similarly, in the following text, the power-on mode is switched at each time t22 to t26 based on the detection of the zero-crossing of the BEMF of the non-powered phase.

[0121] Detailed operation of mutual inductance detection driver

[0122] Figure 7 It is shown Figure 5 The flowchart shows the detailed operation of the mutual inductance detection drive in step S60.

[0123] exist Figure 7 In step S200, MCU 40 determines whether the time Tkick elapsed since the start of the gear shift (referred to as the shift time Tkick) exceeds a predetermined minimum shift time Tmin. If the shift time Tkick exceeds the minimum shift time Tmin (as in step S200), then MCU 40 proceeds the process to step S210.

[0124] In step S210, the MCU 40 detects the induced voltage A of the non-energized phase at the moment the PWM pulse is turned on. The MCU 40 then detects the induced voltage B of the non-energized phase at the moment the subsequent PWM pulse is turned off. Variable C is the value obtained by subtracting induced voltage B from induced voltage A, corresponding to the induced voltage based on mutual inductance. Furthermore, when the rotor rotates, each of the induced voltages A and B also includes BEMF. BEMF can be eliminated by subtracting induced voltage B from induced voltage A.

[0125] More precisely, consider the offset value center_m of the differential amplifier 63 near half of the external input voltage VM and the offset value center_d of the differential amplifier 63 near the ground voltage GND. In step S220, the MCU 40 calculates the variable C using formula (6).

[0126] (A-center_m)-(B-center_d)(6)

[0127] The MCU 40 stores the calculated C in memory.

[0128] In the next step S230, MCU 40 stores the absolute value of (A-center_m) as variable D in memory. Instead of the absolute value of (A-center_m), the absolute value of (B-center_d) can be stored as variable D in memory. When the induced voltage based on mutual inductance (i.e., variable C) is 0, the absolute value of either (A-center_m) or (B-center_d) will be equal to the absolute value of BEMF.

[0129] In the next step S240, MCU 40 determines whether the power-on mode is a, c, or e, or b, d, or f. As an example, if the rotor's rotation direction is... Figure 3 If the rotor's rotation direction is CW and the energizing mode is b, d, or f, then MCU 40 proceeds to step S250. If the rotor's rotation direction is CW and the energizing mode is a, c, or e, then MCU 40 proceeds to step S260. If the rotor's rotation direction is... Figure 3 If the CCW direction described above is b, d, or f, then MCU 40 proceeds the process to step S260. If the rotor's rotation direction is the CCW direction and the energizing mode is a, c, or e, then MCU 40 proceeds the process to step S250. Since the CW and CCW directions are for ease of explanation, the process can be performed in the opposite direction to the above.

[0130] First, the process proceeds to step S250. When the variable C calculated by the above formula (6) is positive at the start of the gear shift, the process proceeds to step S250. In step S250, the MCU 40 determines whether the variable C is greater than the threshold voltage + margin1.

[0131] If variable C is greater than +margin1 (in step S250), then MCU 40 proceeds to step S310. In step S310, if the shift time Tkick is greater than the maximum shift time Tmax (in step S310), then MCU 40 stops shifting and determines that it has timed out (step S320). On the other hand, in step S310, if the shift time Tkick is equal to or less than the maximum trigger time Tmax, then MCU 40 returns to step S210.

[0132] In step S250, if variable C is equal to or less than the margin1 value (no in step S250), then MCU 40 proceeds to step S270. In step S270, MCU 40 determines whether the shift time Tkick is equal to or less than the target shift time Tspeed. When the shift time Tkick exceeds the target shift time Tspeed (no in step S270), MCU 40 determines that the engine speed is insufficient (step S300). In this case, MCU 40 returns to the previous step. Figure 5 The S50 then initiated gear shifting in energized mode.

[0133] On the other hand, when the shift time Tkick in step S270 is equal to or less than the target shift time Tspeed (Yes in step S270), the MCU 40 proceeds to step S280. In the next step S280, the MCU 40 determines whether the variable D stored in step S230 is equal to or greater than a threshold. Variable D being equal to or greater than the threshold corresponds to the absolute value of BEMF being equal to or greater than the threshold voltage Vth. When variable D is equal to or greater than the threshold (Yes in step S280), the MCU 40 determines that variable D has reached the specified speed (step S300). In this case, the MCU proceeds to step S280. Figure 5 BEMF detection driver in S70.

[0134] Next, the case where the process proceeds to step S260 based on the determination in step S240 will be described. In this case, the variable C calculated by the above equation (6) corresponds to the case where the shift start time is negative. In step S260, MCU 40 determines whether variable C is less than the threshold voltage -margin1. If variable C is less than -margin1 (yes in step S260), then MCU 40 proceeds the process to step S310. Since the subsequent processing has already been described in step S250, it will not be repeated. If variable C is greater than or equal to the value of -margin1 (no in step S260), then MCU 40 proceeds the process to step S270. Since the subsequent processing has already been described in step S250, it will not be repeated.

[0135] In the above scenario, the MCU can determine whether the absolute value of variable C is equal to or less than the threshold voltage of +margin1, without dividing the processing based on the power-on mode. In this case, instead of steps S240, S250, and S260, a step is provided to determine whether the absolute value of variable C is equal to or less than +margin1. Specifically, the MCU 40 proceeds to step S310 when the absolute value of variable C is greater than +margin1, and proceeds to step S270 when the absolute value of variable C is equal to or less than +margin1.

[0136] Example of mutual inductance detection driver

[0137] Figure 8A and Figure 8B In the mutual inductance detection driver Figure 1 An example of the output waveform of a differential amplifier. Figure 8A and Figure 8B In the case of energized mode b, the voltage waveform between the output node TV of phase V (which is a non-energized phase) and the virtual neutral point 72, as well as the waveform of phase U current, are shown. Figure 8A This shows the case where the rotor is not rotating. Figure 8B The diagram illustrates the rotor's rotation. The horizontal axis of each graph corresponds to the rotor's position. That is, Figure 8A and Figure 8B The diagram shows the synthesized output waveforms with different rotor positions. The further to the right, the more the output waveforms are shown at positions where the rotor rotates more in the CW direction.

[0138] Reference Figure 8A Voltages A' and B' are detected. Voltage A' is the voltage when the voltage is applied in the PWM control state, and voltage B' is the voltage when the regenerative current flows immediately afterward. According to the above formula (6), variable C is calculated by formula (7).

[0139] C=(A'-center_m)-(B'-center_d) (7)

[0140] By detecting the positive and negative changes of variable C, the rotor position can be detected at every 60 electrical degrees. Point γ corresponds to when variable C = 0.

[0141] Reference Figure 8B The voltage A during the voltage application state in PWM control and the voltage B immediately following the voltage application state when the regenerative current flows are detected. According to the above formula (6), variable C is calculated using formula (8):

[0142] C=(A-center_m)-(B-center_d) (8)

[0143] Here, in Figure 8B In the voltage waveform during rotor rotation, it can be considered that... Figure 8A BEMF is added to the voltage waveform when the rotor is not rotating. Therefore, the above formula (8) is rewritten as follows.

[0144] C=(A'+BEMF-center_m)-(B'+BEMF-center_d) (9)

[0145] Since BEMF is eliminated, the above formula (9) becomes the same as the above formula (7). That is, the value of variable C when the rotor stops is equal to the value of variable C when the rotor is rotating. Therefore, by determining the polarity of variable C, the rotor position can be detected with the same accuracy as when the rotor is rotating, even when the rotor stops. When variable C is 0, formula (10) is established.

[0146] A'-center_m=B'-center_d=0 (10)

[0147] Therefore, the absolute value of (A-center_m) or (B-center_d) calculated as variable D above is equal to the absolute value of BEMF. In other words, the peak level of BEMF can be detected.

[0148] exist Figure 8A and Figure 8B In this context, margin1 is a threshold used to determine the polarity of C. That is, in determining the polarity of the actual variable C, in... Figure 8A and Figure 8B The rotor position is detected at the position before point γ in the waveform. Since the BEMF waveform is close to a sine wave, the detection accuracy of BEMF is not affected even if the variable C = 0, even if it deviates from the rotor position by an electrical angle of about 30 degrees.

[0149] Figure 9A , Figure 9B and Figure 9C It is used to explain in Figure 8A and Figure 8B A diagram showing the rotor positions at points α, β, and γ. Figure 9A , Figure 9B and Figure 9C The diagram shows a cross-sectional view of a brushless DC motor 30 with a pole pair number of 2. Figure 9A It shows Figure 8A and 8B The case of point α, Figure 9B It shows Figure 8A and 8B The case of the mid-β point, Figure 9C It shows Figure 8A and 8B The case at point γ. In energized mode b, the winding current flows from phase W to phase U, and the voltage of phase V, which is the non-energized phase, is detected.

[0150] like Figure 9A As shown, point α represents the rotor position 120 electrical degrees prior to point γ. In this case, phase V generates a positive voltage.

[0151] like Figure 9B As shown, point β represents the rotor position 90 electrical degrees prior to point γ. In this case, the BEMF detected in phase V is approximately 0. Figure 9C As shown, point γ represents the rotor position when variable C (i.e., the induced voltage based on mutual inductance) is 0.

[0152] Effect

[0153] As described above, the brushless DC motor 30 according to this disclosure detects the rotor position by utilizing the induced voltage generated by mutual inductance in the non-energized phase based on the current magnetic field of the energized phase. Therefore, before sensorless control based on BEMF zero-crossing detection becomes feasible, the rotor position can be detected during the initial startup of the brushless DC motor 30.

[0154] Here, by obtaining the difference between the induced voltage of the non-energized phase during the voltage application period of PWM control and the induced voltage of the non-energized phase immediately after or before the induced voltage during the regeneration period, the rotor position can be detected without being affected by BEMF. Therefore, the same determination threshold can be used to detect the moment when the difference is 0, whether the rotor is not rotating or rotating.

[0155] Furthermore, when the aforementioned difference approaches zero, the amplitude of BEMF can be detected from the mutual inductance voltage. Previously, when the PWM frequency was low, a large constant CR filter was required to detect BEMF; however, this embodiment can detect it with high accuracy without using a CR filter.

[0156] Although the present invention has been specifically described based on the embodiments, the present invention is not limited to the above embodiments, and needless to say, various modifications can be made without departing from its spirit.

Claims

1. A semiconductor device for controlling a three-phase motor, comprising: A detection circuit is coupled to an output node corresponding to each phase of the three-phase motor, and the detection circuit is configured to detect the voltage generated in the output node of the non-energized phase of the three-phase motor. as well as Controller When starting the three-phase motor, the controller applies a drive voltage to any two phases of the motor based on an estimated magnetic pole position of the rotor in a stopped state. The controller also estimates the rotor position based on the voltage difference between the voltage detected by the detection circuit during the drive voltage application period and the voltage detected by the detection circuit during a regeneration period immediately following or before the drive voltage application period. Specifically, when the absolute value of the differential voltage is less than or equal to a first threshold, the controller switches the phase to which the driving voltage is applied.

2. The semiconductor device of claim 1, wherein the driving voltage is based on pulse width modulation.

3. The semiconductor device according to claim 1, wherein, When the absolute value of the differential voltage is equal to or less than the first threshold, the controller estimates the absolute value of the voltage detected by the detection circuit as the absolute value of the back electromotive force.

4. The semiconductor device according to claim 3, wherein, When the estimated absolute value of the back electromotive force is greater than or equal to the second threshold, the controller switches to a drive mode that controls the three-phase motor based on the zero-crossing point of the voltage detected by the detection circuit.

5. The semiconductor device according to claim 3, wherein, When the estimated absolute value of the back electromotive force is greater than or equal to a second threshold, and the voltage application time from the start time of applying the drive voltage to two phases of the three-phase motor is less than or equal to a threshold time, the controller switches to a drive mode that controls the three-phase motor based on the zero-crossing point of the voltage detected by the detection circuit.

6. The semiconductor device according to claim 3, wherein, The controller stops applying the drive voltage when the voltage application time from the start time of applying the drive voltage to two phases of the three-phase motor exceeds a threshold time, and when the absolute value of the differential voltage is not less than or equal to the first threshold.

7. The semiconductor device of claim 1, wherein the controller estimates the magnetic pole position of the rotor in the stopped state by inductive sensing.

8. The semiconductor device according to claim 1, wherein the detection circuit further comprises: A switching circuit, configured to select at least one phase of the three phases, and A differential amplifier configured to amplify the difference between the voltage of the output node of the phase selected by the switching circuit and the voltage of the virtual neutral point.

9. The semiconductor device of claim 1, wherein the controller controls the inverter circuit such that the regenerative current flows through the lower arm but not through the upper arm during the regeneration period.

10. A motor drive system, comprising: Three-phase motor; An inverter circuit configured to drive the three-phase motor, and including an output node corresponding to each phase of the three-phase motor; A detection circuit coupled to the output node and configured to detect the voltage generated in the output node of the non-energized phase of the three phases; as well as Controller When starting the three-phase motor, the controller applies a drive voltage to any two phases of the three-phase motor via the inverter circuit based on the estimated magnetic pole position of the rotor in a stopped state. The controller also estimates the rotor position based on the voltage difference between the voltage detected by the detection circuit during the drive voltage application period and the voltage detected by the detection circuit during a regeneration period immediately after or before the drive voltage application period. Specifically, when the absolute value of the differential voltage is less than or equal to a first threshold, the controller switches the phase to which the driving voltage is applied.

11. A method for starting a motor, comprising: Estimate the magnetic pole position of the rotor when it is stationary during the start-up of a three-phase motor; Based on the estimated magnetic pole position of the rotor in the stopped state, a drive voltage is applied to any two phases of the three-phase motor; The position of the rotor is estimated based on the voltage difference between the induced voltage generated in the non-energized phase during the driving voltage application period and the induced voltage generated in the non-energized phase during the regeneration period immediately after or before the driving voltage application period. as well as When the absolute value of the differential voltage is less than or equal to the first threshold, the phase to which the driving voltage is applied is switched.

12. The motor starting method according to claim 11, wherein the driving voltage is based on pulse width modulation.

13. The motor starting method according to claim 11, further comprising: The absolute value of the induced voltage is estimated as the absolute value of the back electromotive force, wherein the induced voltage is the voltage generated in the non-energized phase when the absolute value of the differential voltage is less than or equal to the first threshold.

14. The motor starting method according to claim 13, further comprising: When the estimated absolute value of the back electromotive force is greater than or equal to the second threshold, the inverter circuit is switched to a drive mode, which controls the inverter circuit based on the zero-crossing point of the induced voltage in the non-energized phase.

15. The motor starting method according to claim 13, further comprising: When the estimated absolute value of the back electromotive force is greater than or equal to a second threshold and when the voltage application time from the start time of applying the drive voltage to two phases of the three-phase motor is equal to or less than the threshold time, the inverter circuit is switched to a drive mode that controls the inverter circuit based on the zero-crossing point of the induced voltage in the non-energized phase.

16. The motor starting method according to claim 13, further comprising: When the voltage application time from the start time of applying the drive voltage to two phases of the three-phase motor exceeds a threshold time and when the absolute value of the differential voltage is not less than or equal to the first threshold, the application of the drive voltage is stopped.

Images