This article discusses sensorless control methods for brushless DC motors . The core of this method is to obtain the rotor position signal. Currently, the main methods commonly used in the market are as follows.
1. Magnetic flux linkage observation method
The motor flux linkage signal is directly related to the rotor position, so the rotor position signal can be determined by the value of the rotor flux linkage. However, the motor rotor flux linkage cannot be directly detected. To obtain the motor rotor flux linkage value, it is necessary to first measure the phase voltage and current of the motor, establish a functional equation that is independent of the rotor speed and directly related to the rotor flux linkage, and calculate the flux linkage value. This method involves a large amount of computation, and the phase voltage and current contain a large amount of interference signals. Accurate measurement also requires high hardware and software costs, so it is rarely used.
2. Back potential zero-crossing detection method
In brushless DC motors, the back EMF of the windings typically alternates between positive and negative. When the back EMF of a certain phase winding crosses zero, the rotor's direct axis coincides with the axis of that phase winding. Therefore, by detecting the zero-crossing points of each back EMF, several key rotor positions can be determined, thus eliminating the need for a rotor position sensor and achieving sensorless brushless DC motor control. This is currently the most widely used sensorless BLDCM control method.
The disadvantage of this method is that the back EMF signal is zero or very small when stationary or at low speed, making it difficult to accurately detect the back EMF of the winding. As a result, an effective rotor position signal cannot be obtained, and the system has poor low-speed performance, requiring an open-loop method for starting. In addition, in order to eliminate the interference signal caused by PWM modulation, the back EMF signal needs to be deeply filtered, which causes a phase shift in the signal related to the motor speed. In order to ensure correct commutation, this phase shift needs to be compensated.
3. Third harmonic integration method of back potential
Since the back electromotive force (EMF) of a BLDCM is a typical trapezoidal wave, containing the fundamental frequency and its higher harmonic components, the third harmonic and its odd-numbered harmonics can be obtained by simply superimposing the three-phase phase voltages of the armature. The third harmonic component of the back EMF can then be extracted and integrated. When the integral value is zero, the switching signal of the power device is obtained.
There are two methods to obtain the third harmonic signal of the back electromotive force (EMF): one method uses the neutral point of the motor and the neutral point of the star resistor connected in parallel to the three-phase winding terminals of the motor to obtain the third harmonic component of the back EMF; for motors without a neutral point, the DC side midpoint voltage and the neutral point of the star resistor network can be used to obtain the third harmonic component of the back EMF. The obtained signal is then filtered to remove the higher-order components of the third harmonic. Since the lowest higher-order components are at least nine times the fundamental frequency, the filter requirements are low. Therefore, it has a wider operating range than direct zero-crossing comparison of the back EMF. This method avoids interference caused by inverter switching, but the amplitude of the third harmonic is smaller than that of the back EMF, making it difficult to detect, especially at low speeds where the third harmonic signal is weaker, making it difficult to obtain the rotor position signal.
4. Freewheeling diode method
This method determines the commutation timing of the motor power transistors by monitoring the conduction status of the freewheeling diodes connected in parallel across the inverter power transistors. Since one phase of the three-phase winding of a BLDCM is always off, the switching sequence of the six power transistors can be obtained by monitoring the conduction and turn-off status of the six freewheeling diodes. This method can improve the motor's speed range, especially by widening the lower limit of the speed range. However, this method requires the inverter to operate in PWM chopping mode, alternating between the upper and lower power transistors, increasing the control difficulty. Furthermore, removing invalid signals from the freewheeling diodes and misleading signals caused by glitches is not easy. This method also suffers from significant detection errors; non-constant back EMF coefficients, winding inductance, and non-standard trapezoidal back EMF waveforms can all cause rotor position errors. Because this method requires a parallel detection circuit on the diodes, it is difficult to implement with integrated power devices (such as IPMs). Due to these drawbacks, this method is not widely used in China, and the technology is relatively immature.
5. Inductance method
There are two forms of inductive methods: one for salient-pole permanent magnet brushless DC motors and the other for permanent magnet brushless DC motors with embedded magnet structures. The first inductive method primarily determines the change in inductance by applying a detection voltage to the motor windings during startup. In a salient-pole motor, the winding self-inductance can be expressed as an even-order cosine function of the angle between the winding axis and the rotor's direct axis. By detecting changes in the winding self-inductance, the approximate position of the rotor axis can be determined, and then the polarity can be determined based on the trend of core saturation, thus ultimately obtaining the correct position signal. This method is more complex and can only be applied to salient-pole motors, so it is rarely used now. Compared to the first method, the second method is the true inductive method. In an embedded permanent magnet brushless DC motor, the winding inductance changes accordingly with the rotor position. By detecting these changes and performing certain calculations, the rotor position signal can be obtained. This method requires continuous real-time detection of the winding inductance, increasing the difficulty of implementation and limiting its widespread application.
6. State Observer Method
The basic idea of the "state observer method" is to use parameters such as motor speed, rotor position angle, and current as state variables. Based on these defined state variables, a mathematical model of the motor is established, and discrete values of the state variables are obtained through digital filtering, thereby achieving motor control. The "state observer method" effectively solves the problem of motor control difficulties under high speed and heavy load conditions, and its good anti-interference capability makes it more suitable for harsh working environments. However, the massive computational load of the "state observer method" limits its application to some extent. This method typically uses a digital signal processor (DSP) to handle the large computational load, thus increasing system cost and making it uncommon in practical applications.
The driving methods of brushless linear motors can be divided into various types, each with its own characteristics.
By driving waveform:
1. Square wave drive: This drive method is easy to implement and facilitates sensorless control of the motor.
2. Sine drive: This driving method can improve motor operation and make the output torque more uniform, but the implementation process is relatively complex. Furthermore, this method has two forms: SPWM and SVPWM (Space Vector PWM), with SVPWM performing better than SPWM.
