Abstract: This paper introduces the SVPWM mechanism, designs the software and hardware scheme of current loop control for servo motors, studies the current loop servo control strategy, analyzes the current sampling principle, and discusses the dead-time compensation problem of inverters in motor drives. Finally, the software implementation method of current loop interruption is described in detail. Keywords : space vector pulse width modulation (SVPWM); inverter; dead-time compensation 0 Introduction In the past ten years or so, DSP controllers have been widely used in motor control. TI's TMS320F2812 DSP features more complete peripheral control interfaces and richer motor control peripheral circuits. Its Event Manager (EV) includes a hardware SVPWM generation circuit. SVPWM generation offers advantages such as simple hardware structure, high control precision, strong real-time performance, and easy software programming. The motor control strategy mainly employs a three-loop control: position loop, speed loop, and current loop. Their main function is to minimize errors, achieve higher, more accurate, and faster control precision, and realize automatic control. The current control loop is an inner loop, and the implementation of the SVPWM control algorithm mainly focuses on the current loop. Within one current loop cycle, it is necessary to complete the sampling of feedback current, voltage dead-zone compensation, and output voltage calculations. Its performance indicators, especially dynamic characteristics, will comprehensively affect the speed loop and position loop, thus affecting the entire servo system. 1 SVPWM and Motor Current Loop Control In motor frequency conversion speed regulation, pulse width modulation technology has been widely used. Compared with the classic pulse width modulation method, the space vector pulse width modulation (SVPWM) method has advantages such as high DC voltage utilization, simple control, low loss, and easy digital implementation, and is widely used in motor control. 1.1 SVPWM Mechanism Space vector PWM refers to a special switching mode of the six power transistors in a three-phase power inverter. The essence of the space vector PWM method is to use eight switching combinations of the six power transistors to give the power supply voltage vector of the motor. Through α-β coordinate transformation, the phase voltages corresponding to the eight state combinations are mapped to the α-β coordinate plane, that is, the three vectors (a, b, c) are vertically mapped to a two-dimensional coordinate (α-β coordinate), so that six non-zero vectors and two zero vectors can be obtained. The six non-zero vectors form a hexagon, the angle between adjacent vectors is 60°, and the two zero vectors are at the origin, as shown in Figure 1. [align=center] Figure 1 Basic voltage space vector diagram in α and β coordinate system[/align] By detecting the direction and magnitude of the motor phase current, the two components of Uout can be calculated. Based on the magnitude of these two components and their respective sector intervals, the duration of Ux's action is obtained and written into the comparison register of the DSP2812's Event Manager (EV). The PWM wave is then generated through EV. 1.2 DSP2812 Current Loop Control Structure The current loop structure is shown in the figure below: Ia+Ib+Ic=0, therefore, only two current paths need to be detected. The motor phase current value is detected by a Hall sensor, and the detected signal Iu is output after IU conversion by an operational amplifier. The SVPWM hardware structure diagram is shown in Figure 3: [align=center] Figure 2 Current Loop Control Structure Diagram[/align] The current loop control algorithm uses PI, SVPWM (Space Vector PWM). PI is the classic proportional-integral control, while SVPWM control technology can significantly reduce the harmonic components of the inverter output current and reduce pulsating torque. It is particularly easy to implement digitally and is widely used in the field of motor control. The current loop control structure diagram is shown in Figure 2. After Clarke transformation, Pake transformation, and inverse Clarke transformation, it acts on the inverter to generate a sine wave. 2. Current Sampling For digital servo motor control systems, the accuracy and real-time performance of current sampling largely determine the dynamic and static performance of the system. Therefore, accurate current detection is an important condition for improving the system's control accuracy, stability, and speed, and is also key to realizing a high-performance closed-loop control system. Since this paper studies a three-phase balanced system, [align=center]Figure 3 SVPWM Hardware Structure Diagram[/align], to reduce high-order harmonics in the current signal and achieve high-performance closed-loop control, in addition to designing a suitable current loop controller, it is also necessary to minimize high-order harmonic components in the current feedback value. The current sampling sensor can be installed at the bottom of each IGBT bridge arm. In a PWM cycle, sampling the current at the beginning or midpoint of each switching cycle of the PWM wave can obtain a fundamental current value with relatively few harmonic components, which is beneficial for achieving high-precision current closed-loop control. 3. Dead Time Compensation of DSP2812 In voltage-type pulse width modulation (PWM) inverter circuits, dead time must be inserted to avoid shoot-through of switching devices on the same bridge arm. However, the dead-time effect is a significant nonlinear factor affecting the inverter's voltage and current output. For motor drive systems, the dead-time effect causes severe distortion of voltage and current at low motor speeds, leading to torque pulsation and harmonics. Therefore, the inverter's dead-time must be compensated. There are two common dead-time compensation methods: hardware compensation and software compensation. Hardware compensation requires additional hardware circuitry, obtaining the necessary compensation voltage signal by comparing the actual output voltage with a voltage reference value. Software compensation uses a purely software-based approach, incorporating a dead-time compensation algorithm into the motor control program. This type of compensation algorithm can also be called dead-time compensation. In the TMS320LF 2812, the dead-time compensation pulse width can be implemented by modifying CMPR1 and CMPR2 in the event manager EV. 4. Software Implementation The flowchart of the current loop control program with dead-time compensation is shown in Figure 4. [align=center] Figure 4 Software Flowchart of Current Loop with Dead-Time Compensation[/align] The statement that enters the current loop timer interrupt from the main program is: PieVectTable.T1PINT = &EvaTimer1P_ISR; System initialization function void InitEv (void), sets the T1 period to 500us, T1 control register, compare enable, timer enable, continuous increment/decrement counting mode, sets PWM output 135 as active high and 246 as active low. void InitEv(void) {…… EvaRegs.T1PR=0x0BEBB; EvaRegs.T1CON.all=0x0842; EvaRegs.T1CMPR = 0x3c00; EvaRegs.ACTRA.all = 0x666; EvaRegs.T1CNT = 0x0000; EvaRegs.DBTCONA.all=0x088f; EvaRegs.COMCONA = 0xA600; ……} The Clarke transform, Pake transform, PI algorithm, and inverse Clarke transform appear as functions, defined as follows: void clarke_calc(CLARKE* g_Ci); void park_calc(PARK* g_Pi); double PID_calc1(PID* g_pid, double ProcessPoint); double PID_calc2(PID* g_pid, double ProcessPoint); void `ipark_calc(IPARK* g_Iv); void CALC_SECTOR(); void CALC_T1T2();` Timer 1 counts to the set period value, and the comparison register is automatically reloaded according to the calculated time value, causing the duty cycle of the output PWM waveform to change. The loading statements are as follows: `EvaRegs.CMPR1=T1+Td1; EvaRegs.CMPR2=T1+T2+Td2; EvaRegs.ACTRA.all&=0x0fff; EvaRegs.ACTRA.all|=(SVDIR << 15) + (vector << 12); ` 5 Conclusion This paper details the software implementation method of SVPWM, implementing full digital control of SVPWM using TI's DSP chip. Through improvements to the current sampling and dead-time compensation algorithm, the influence of motor dead time is effectively reduced, harmonics are eliminated, and current output capability is improved. The designed current loop has higher control efficiency. The dynamic performance has been improved, laying a good foundation for speed control and position loop control in AC servo systems. References [1] Chen Rong, Yan Yangguang. Research on control strategy of AC permanent magnet servo system [J]. Journal of Electrical Machines and Control, 2004, (3): 205-208. [2] Zhang Ximing, Chen Liming, Ni Guangzheng. An electric drive system for electric vehicles based on TMS320F2812. Mechanical and Electrical Engineering, 2008, 25 (3): 61-63. [3] Liu Liang, Deng Minggao, Ouyang Honglin, et al. Research on dead zone compensation method of PWM inverter based on predictive current control [J]. Journal of Electrical Engineering, 2005, 8 (20): 78-98. [4] Xu Jiayan, Xu Guoqing, Kang Jinsong. SVPWM dead zone analysis and compensation based on permanent magnet synchronous motor. Electrical Drive, 2007, 37 (02): 29-31 [5] Wang Chengyuan, Xia Jiakuan, Yang Junyou, Sun Yibiao. Modern Control Technology of Electric Motors [M]. Beijing: Machinery Industry Press, 2006. About the author: Sun Jie, the first author, is a female born in 1983. She graduated from the Department of Automation of China University of Petroleum in 2006 with a bachelor's degree. Her technical title and position are (/currently a master's student in Mechanical Design and Theory at Beijing Information Science and Technology University); she is mainly engaged in the research of motor control and robot technology.