Discussion on key technologies of BLDC position sensorless control

01Overview:

In brushless DC motor control systems, position sensors (such as Hall sensors, etc.) provide the most direct and effective detection method for rotor position, but they also make the motor larger, require more signal leads, and increase production costs. In some applications (such as high temperature and high pressure), the unreliability of position sensors brings the risk of system failure. Therefore, people are committed to finding a control method for brushless DC motors without position sensors. This article will discuss the key technologies of position sensorless control, including motor drive methods, PWM modulation methods, and rotor position detection methods.

02Motor drive mode selection:

1. Analysis of main power circuit driving mode

Brushless DC motors can have multi-phase structures, each of which can be driven by a full-bridge or half-bridge circuit, and full-bridge drive can be divided into star and delta connections and different power-on methods. Different choices will result in different performance and costs for the motor and control system. Taking the most widely used three-phase brushless DC motor as an example, there are three-phase half-bridge drive, three-phase star full-bridge drive, three-phase delta full-bridge drive and other methods as shown in Figure 1 below:

(a) Half-bridge drive mode

(b) Half-bridge drive mode

Figure 1: Schematic diagram of brushless DC motor drive method

As shown in Figure 1 (a) above, the three-phase half-bridge drive circuit is simple, but the utilization rate of the motor winding is very low. Each winding is only energized for 1/3 of the cycle, and the other 2/3 of the time is in a power-off state. The winding is not fully utilized, and the torque fluctuates greatly during operation. For occasions with higher requirements, a three-phase full-bridge circuit is generally used, as shown in Figure 1 (b) above.

Regardless of the connection method used for the motor winding, the three-phase full-bridge drive circuit has two power-on modes: two-by-two conduction and three-by-three conduction. The two-by-two power-on mode means that two switch tubes are turned on or modulated at every moment, and the phase is changed every 60 degrees of electrical angle. Each phase change changes the state of a switch tube, and each switch tube is turned on for 120 degrees of electrical angle; the three-by-three power-on mode means that three switch tubes are turned on or modulated at the same time at every moment, and the phase is changed every 60 degrees of electrical angle, and each switch tube is powered on for 180 degrees of electrical angle. However, in the three-by-three power-on mode, there are strict regulations on the order of turning off and on the switch tubes. If you are not careful, the upper and lower bridge arms will be turned on at the same time, causing the DC power supply to short-circuit and burn out.

Based on the above analysis, this paper adopts a three-phase star full-bridge drive circuit and a two-by-two conduction power-on method to explore the key technologies of position sensorless control.

2. Six-step commutation method

After the brushless DC motor adopts the three-phase star full-bridge driving method with two-phase power supply, the phase is changed six times in each electrical cycle, which is what we often call the six-step phase change method. According to the different energized windings, an electrical cycle is evenly divided into 6 steps, called 6 intervals or 6 states. The phase change occurs at the switching moment of two adjacent states and is completed by the switching of the switch tube. The principle of the six-step phase change method is shown in Figure 2 below.

(a) Current direction corresponding to each state of six-step commutation

(b) Stator winding back electromotive force waveform and switch tube conduction sequence

Figure 2: Schematic diagram of six-step commutation principle

Figure 2 (a) shows the direction of current flowing through the motor winding in each step of the six-step commutation, and Figure 2 (b) shows the back electromotive force waveform of the motor winding and the conduction status of the switch tube in each step. The conduction order of each switch tube is V1V4, V1V6, V3V6, V3V2, V5V2, V5V4, V1V4... When V1 and V4 are turned on, the current flows from V1 into the A-phase winding, then flows out from the B-phase winding, and flows back to the power supply through V4. In this state, the C-phase winding is not energized, that is, it is in a suspended state. In each state, two phases of the winding are energized and the other phase of the winding is suspended. This is an important feature of the six-step commutation method. The position sensorless control that we will discuss in this article is based on this realization.

03PWM modulation mode:

PWM control is the most commonly used motor speed regulation method, especially with the development of power electronic devices such as IGBT and MOSFET in recent years, the modulation frequency of PWM can reach tens or even hundreds of kHz, which provides conditions for the motor's wide speed, fast response and flexible speed regulation. PWM control mainly modulates the switching state of the bridge inverter bridge power tube through PWM waves to control and regulate the current. According to the different PWM action time and the switch tube that acts, PWM modulation can be divided into five modes. Within the 120-degree electrical angle time when each switch tube is turned on, the five modulation modes are shown in Figure 3 below.

Figure 3: Five PWM modulation methods under 120-degree conduction mode

(1) H_PWM-L_PWM mode: The upper and lower arms of the inverter bridge are modulated using complementary PWM signals;

(2) ON_PWM mode: In the 120-degree electrical angle conduction space of each switch tube, the first 60 degrees electrical angle remains constant, and the last 60 degrees electrical angle is PWM modulated;

(3) PWM_ON mode: In the 120-degree electrical angle conduction space of each switch tube, the first 60 degrees electrical angle is PWM modulated, and the last 60 degrees electrical angle remains constant;

(4) H_PWM-L_ON mode: In each power-on state, the switch tube in the upper bridge arm of the inverter bridge adopts PWM modulation, and the switch tube in the lower bridge arm remains constantly on;

(5) H_ON-L_PWM mode: In each power-on state, the switch tube in the upper bridge arm of the inverter bridge remains constantly on, and the switch tube in the lower bridge arm adopts PWM modulation.

Among the five modulation modes, the mode of modulating the upper and lower bridge arms at the same time, such as H_PWM-L_PWM, is called the "full chopping" modulation mode; the other four modulation modes are called "half chopping" modulation modes. The switching loss and current pulsation of the stator winding in the "full chopping" mode are twice that of the other "half chopping" modes, and in the four modulation modes of "half chopping", during the upper bridge commutation process, the torque pulsation under PWM_ON mode and H_PWM-L_ON is smaller than that under ON_PWM mode and H_ON-L_PWM mode; during the lower bridge commutation process, the torque pulsation under PWM_ON mode and H_ON-L_PWM is smaller than that under ON_PWM mode and H_PWM-L_ON mode.

Considering the simplicity of control, we choose the most commonly used H_PWM-L_ON mode (also known as upper bridge chopping and lower bridge constant on) in this article, that is, in each power-on state, only the upper bridge arm is PWM modulated, and the lower bridge arm remains constant on. Taking state 1 as an example, AB phase is turned on. When PWM is high, V1 and V4 are turned on, and the power passes through V1 and V4, and the current increases; when PWM is low, V1 is turned off, V4 is turned on, and the current continues through the diode. The use of H_PWM-L_ON mode can effectively reduce the torque pulsation of the motor, especially at high speeds. The complete PWM control signal is shown in Figure 4 below.

Figure 4: PWM control signal waveform

04 Implementation of the back electromotive force zero-crossing detection method:

For a brushless DC motor with a trapezoidal back electromotive force, the zero-crossing point of the suspended back electromotive force voltage can be obtained by detecting the zero-crossing point of the suspended phase voltage. However, the lead wires of the motor generally only have the leads of the three-phase windings A, B, and C, and the physical quantities that can be directly detected are only the terminal voltage and phase current. Therefore, only by processing and calculating these physical quantities can the back electromotive force of the motor be obtained and its zero-crossing point be detected.

Since the neutral point of most motors is not led out, it is impossible to directly compare the stator terminal voltage with the neutral point voltage to obtain the zero crossing point. In view of this situation, one solution is to compare the terminal voltage with half of the DC bus voltage, assuming that the back EMF zero crossing event occurs when the terminal voltage is equal to VDC/2, as shown in Figure 5 below. This circuit is easy to implement, and only requires a comparator to be connected to the winding lead wire, so a total of three comparators are required. However, the terminal voltage signal detected by this method has positive and negative phase shifts, and in most cases the rated voltage of the motor is less than the VDC voltage, so the back EMF zero crossing event does not always occur at VDC/2, so the detection is inaccurate.

Figure 5: Comparison diagram of terminal voltage and half of DC bus voltage

Another method is to use a resistor divider network to form a virtual neutral voltage at the three-phase stator terminal voltage, and obtain the back electromotive force zero-crossing point by comparing the terminal voltage with the virtual neutral point voltage, as shown in Figure 6 below. However, since the motor uses PWM speed regulation, high-frequency interference will be superimposed on the stator terminal voltage, affecting the acquisition of the back electromotive force zero-crossing point. In many cases, it is achieved by using resistor divider and RC low-pass filtering, but this will cause the back electromotive force signal to be greatly attenuated and will bring about the phase shift problem of the zero-crossing point. Phase compensation must be performed later, which increases the complexity of control.