Speed ​​sensorless controller for small power permanent magnet synchronous motor

　　I. Preface

　　In traditional AC vector conversion control systems, speed sensors are indispensable. For ordinary AC motors, speed sensors have three functions: first, to obtain speed feedback signals to achieve closed-loop speed control; second, to add the slip angular frequency to obtain the stator current angular frequency given value for frequency control; third, to observe the rotor flux with the current model in the low-speed range for field-oriented control. If vector control of PMSM is to be implemented so that the direction of the stator current is orthogonal to the direction of the magnetic flux generated by the permanent magnet in space, a position sensor is also required to determine the position of the rotor poles. Based on the position information, the control circuit supplies power to the three-phase stator winding with the correct phase and phase sequence, and generates a constant torque through the alternating stator current, thereby achieving precise control of the system. Most motor speed and pole position detection uses mechanical sensors such as photoelectric encoders or rotary transformers. In practical applications, there are the following problems:

　　(1) High-precision speed sensors are expensive, which significantly increases the cost of system development (including sensors and electronic circuits) for some small-capacity devices.

　　(2) The speed sensor is difficult to install and there is a concentricity problem. This problem can only be solved satisfactorily in specially processed motors. In general motors, due to installation problems, the speed sensor often becomes a source of system failure, greatly reducing the mechanical robustness of the system and bringing difficulties to maintenance.

　　(III) The transmission of speed signal is limited by distance. If the distance is long, it will bring many interference signals. Therefore, it is limited to some high-performance occasions, which greatly limits the scope of application:

　　(iv) When selecting a frequency converter, the parameters of the speed sensor must be taken into account to make them match and have poor interchangeability;

　　(V) In harsh environments, since most sensors have weak internal and output signals and poor anti-electromagnetic interference capabilities, the effects of temperature, humidity, vibration, etc. will cause the performance of the sensor to be unstable, restricting the accuracy of detection. If a rotary transformer is used, the speed and magnetic pole position information is obtained through signal processing, and the demodulation process is more complicated, increasing the complexity of hardware and software and the difficulty of control strategies:

　　(VI) All sensors will generate a certain degree of static and dynamic friction on the motor drive shaft, and also add a certain inertia to the motor shaft.

　　The permanent magnet synchronous motor (PMSM) made of the third-generation permanent magnet material - neodymium iron boron (NdFeB) has significant advantages such as small size, light weight, low loss, high efficiency, and flexible motor shape and size. It is increasingly widely used in industrial and agricultural production, aerospace, national defense and daily life. At the same time, my country is rich in rare earth resources. Therefore, researching and developing high-efficiency rare earth permanent magnet synchronous motors and transforming the export of rare earth resources into the export of high-value-added rare earth permanent magnet motor products can promote the product structure adjustment and upgrading of the motor industry and the fan and water pump industry, thereby creating huge economic benefits. Therefore, the research and development of permanent magnet synchronous motor controllers has great application value.

　　2. Implementation Method

　　Introduction to AVR microcontroller and IRMCF341 chip

　　This solution is to be implemented using ATMEL's ATmega64 microcontroller and IRMCF341, a high-performance sensorless sinusoidal motor control IC for home appliances.

　　ATmega64:

　　High reliability, powerful functions, high speed, low power consumption and low price have always been important indicators for measuring the performance of microcontrollers, and are also necessary conditions for microcontrollers to occupy the market and survive.

　　The early single-chip microcomputers adopted a safe solution mainly due to the low level of technology and design, high power consumption and poor anti-interference performance: using a higher frequency division coefficient to divide the clock, which makes the instruction cycle long and the execution speed slow. Although the later CMOS single-chip microcomputers adopted measures such as increasing the clock frequency and reducing the frequency division coefficient, this situation has not been completely changed (51 and 51 compatible). Although some reduced instruction set microcontrollers (RISC) have been introduced, they still follow the practice of clock division.

　　The introduction of AVR microcontrollers completely broke this old design pattern, abolished the machine cycle, abandoned the practice of complex instruction computers (CISC) to pursue complete instructions; adopted a reduced instruction set, using words as the unit of instruction length, and arranged rich operands and operation codes in one word (most of the single-cycle instructions in the instruction set are like this), with a short instruction fetch cycle, and can pre-fetch instructions to achieve pipeline operation, so instructions can be executed at high speed. Of course, this speed jump is backed by high reliability.

　　The Harvard-structured RISC microcontroller AVR developed by Atmel in 1997 absorbed the advantages of PIC and 8051 series microcontrollers, and also made some major improvements in the internal structure. Its advantages are mainly as follows:

　　The built-in high-quality FLASH program memory can be repeatedly erased and written, supports ISP and IAP, and is convenient for product debugging, development, production, and update. The built-in long-life EEPROM can store key data for a long time to avoid power failure. The large-capacity RAM on the chip can not only meet the use of general occasions, but also more effectively support the use of high-level languages ​​to develop system programs.

　　High speed, low power consumption, with SLEEP function. The execution cycle of an instruction of AVR can reach 50ns, and the power consumption is between 1uA and 2.5mA. AVR adopts Harvard structure and the pre-fetch instruction function of the first-level pipeline, that is, different data buses are used for program reading and data operation. Therefore, when a certain instruction is executed, the next instruction is taken out from the program memory in advance, which allows the instruction to be executed in every clock cycle.

　　Rich peripherals. The AVR microcontroller contains peripherals such as SPI, EEPROM, RTC, watchdog timer, ADC, PWN and on-chip oscillator, which can truly achieve single chip.

　　Good anti-interference performance. With watchdog timer (WDT) safety protection, it can prevent the program from running away and improve the anti-interference performance of the product. In addition, the power supply anti-interference performance is also very strong.

　　Highly confidential. The FLASH that can be burned multiple times has multiple password protection lock (LOCK) functions, so it can be commercialized quickly at a low price, and the program can be changed multiple times (product upgrade), which is convenient for system debugging, and there is no need to waste IC or circuit boards, which greatly improves product quality and competitiveness.

　　Strong driving capability. With large current: 10~20mA (output current) or 40mA (sink current), it can directly drive LED, SSR or relay.

　　Low power consumption. It has six sleep modes and can quickly wake up from low power consumption modes.

　　Ultra-functional reduced instructions. It has 32 general working registers (equivalent to 32 accumulators of 8051), overcoming the bottleneck phenomenon caused by single accumulator data processing.

　　Rich interrupt vectors. It has 34 interrupt sources, and different interrupt vector entry addresses are different, which can respond to interrupts quickly.

　　High reliability. AVR microcontroller has a power-on start counter inside. When the system is reset and powered on by RESSET, the internal RC watchdog timer can delay the MCU to start executing the program after the system power supply and external circuits are stable, which improves the reliability of the system and saves the external reset delay circuit. In addition, the built-in power-on reset (POR) and power drop detection (BOD) also effectively improve the reliability of the microcontroller.

　　IRMCF341:

　　IRMCF341 is a new sinusoidal single-chip control integrated circuit for permanent magnet AC motor applications in household appliances launched by IR. Unlike traditional single-chip microcomputers or DSP solutions, IRMCF341 provides a complete sensorless closed-loop control algorithm for permanent magnet synchronous motors through a proprietary motor control engine (MCETM). The motor control engine (MCETM) contains all motor control elements, motor peripherals, a proprietary motion control timing generator, and a dual-port RAM for exchanging data. IRMCF341 also has a set of proprietary analog-to-digital hybrid circuits to implement single-resistance current sampling and motor current reconstruction algorithms. Only one resistor on the DC negative bus is required to complete motor current sampling and reconstruction, which greatly simplifies system design and reduces system costs.

　　Another feature of IRMCF341 is that users do not need to write motor control programs. Users can use a dedicated graphical compiler to build their own motor control system (usually a custom speed loop function) in the MATLAB/SimulinkTM environment in a building block manner. IRMCF341 also has an embedded high-speed 8-bit 8051 core. Users can flexibly use 8051 programming to implement timing control, user interface, host communication, and upper-level control tasks that are actually required by the system. The 8051 core can be simulated and debugged through the JTAG port (FS2's ISA-M8051EW simulator or other simulators that support Mentor's M8051EW core can be used). The figure below is a typical system structure block diagram based on IRCF341.

　　IRMCF341 is a version for the development stage. Its 48K program memory is RAM, which can be easily accessed from the outside.

　　The EEPROM is loaded with the 8051 and MCE control code. For mass production, either an OTP version or a mask version with identical pinouts can be used.

　　3. Design content

　　This topic studies three aspects:

　　(I) Motor speed estimation algorithm

　　This is the most critical part of this design. The method of estimating the motor speed will be studied. How to design a practical and efficient speed estimation algorithm will be the core research content of this topic.

　　The estimation algorithm used in this design is the flux calculation algorithm, which uses the terminal voltage, terminal current and other parameters of the synchronous motor to directly or indirectly calculate the speed through the ideal mathematical model of the motor.

　　Before introducing the algorithm in detail, the following two problems need to be solved: observation of flux vector and error control of motor parameters.

　　1. Observation of magnetic flux vector

　　In order to correctly calculate the speed, the observation of the magnetic flux vector is essential. This design uses the voltage, current, speed signal and motor parameters of the motor to calculate the magnetic flux vector after different digital-analog operations. Generally, two types of mathematical models are used:

　　2. Error control of motor parameters

　　The premise of using this method to measure the motor speed is that the motor parameters must be set correctly. If the set value changes, it will affect the accuracy of the estimated value. Among the parameters, the most influential are the stator resistance and rotor resistance, which change with the change of the motor temperature.

　　Theoretically, both stator resistance and rotor resistance can be observed by adaptive observers together with speed. However, according to the formula, in steady state, rotor resistance is expressed as Rr/S (S is slip rate), and its error is only related to speed. Therefore, it is not practical to perform adaptive observation of rotor resistance in a speed sensorless system. Generally, the following formula can be used for adaptive adjustment of stator resistance, and rotor resistance can only be corrected proportionally by considering a temperature change function similar to stator resistance.

　　C: System unit circuit

　　a: Phase voltage detection

　　Phase voltage detection consists of three parts: voltage divider resistor, voltage detection (insulated voltage sensor), and filter circuit. Since the main circuit is high voltage, it is necessary to pay attention to the insulation between the main circuit and the control circuit, and select a voltage sensor with good insulation performance.

　　b: Phase current detection

　　Phase current detection uses a Hall-type current sensor, and a specified resistor is connected to its secondary side as a load, so that the output voltage and the measured current can change linearly.

　　Self-detection of parameters of speed sensorless vector control system

　　A. Overview of Parameter Self-Detection

　　The so-called parameter self-detection is to use the inherent hardware resources of the speed control system itself (such as PWM inverter, computer control system) to achieve the purpose of motor parameter determination by executing a series of routine subroutines. In modern AC speed control systems, advanced digital controllers have great flexibility, which is mainly reflected in the flexibility to configure software for different purposes. Under the condition of meeting certain external constraints, it can control the power AC to output various voltage and current test signals, and then after processing the sampled data, it can obtain the motor parameter value with high credibility. It can be seen that parameter self-detection not only expands the function of the speed control system, but also lays the foundation for fundamentally improving the performance of the AC speed control system because it can provide high-precision motor parameters to the control system.

　　B. Offline automatic setting of motor parameters

　　There are three methods for offline self-setting of motor parameters: a. No-load test and locked-rotor test. This is the most classical test method; b. Adaptive algorithm. Use the flux observer of adaptive control to automatically set the motor parameters; c: Use the transition response waveform for self-detection. Motor parameters often vary greatly under different test conditions. In order to improve the measurement accuracy of motor parameters, the parameters should be automatically set as close to the actual operating conditions as possible.

　　C. Implementation and steps of self-test

　　The signals collected by the self-test are only the command current, and the motor speed as well as the step response of and. The specific test items are: 1. stator resistance; 2. slip coefficient; 3. torque coefficient and other motor parameters

　　D. Motor parameters online self-correction

　　As the load changes during operation, the motor temperature rises, causing the rotor resistance to change within a wider range, which will affect the torque change. Therefore, the motor parameters must be corrected online.