A complete analysis of the design of a high-performance fully digital sine wave inverter

This paper proposes a design scheme for a high-performance all-digital sine wave inverter power supply. The scheme is divided into two stages, the front stage uses a push-pull boost circuit to boost the input DC power to a bus voltage of about 350 V, and the back stage uses a full-bridge inverter circuit. After the inverter bridge output is filtered by a filter, an isolation transformer is used for voltage sampling, and a current transformer is used for current sampling to form a feedback link to increase the stability of the power supply output. The boost stage PWM drive and the inverter stage SPWM drive are both generated by the STM32 microcontroller, which reduces hardware expenses. The 400W prototype based on the above scheme has output short-circuit protection, over-current protection, input over-voltage protection, and under-voltage protection. The frequency deviation is less than 0.05Hz at 50Hz output, the full load (400W) efficiency is higher than 87%, the voltage accuracy is 220V±1%, and the THD is less than 1.5%.

Inverter power supplies are widely used, especially precision instruments, which have higher performance requirements for inverter power supplies. High-performance inverter power supplies not only require stable operation, high inverter efficiency, good output waveform characteristics, and complete protection functions, but also require the inverter power supply to be miniaturized, intelligent, and scalable. This paper proposes a pure digital sinusoidal inverter power supply based on the STM32 series microcontroller STM32F103VE. All functions of the power supply are controlled by the microcontroller, and it has the characteristics of stable output voltage and frequency, high efficiency, and complete protection functions.

System Design

The overall framework of the system is shown in Figure 1. The system adopts a high-frequency inverter solution, that is, a structure of front-stage boost plus rear-stage inverter, which can avoid the use of bulky power frequency transformers, effectively reduce the size, weight and cost of the power supply , and improve the efficiency of the power supply. The working principle of the circuit is that the 12V DC input voltage is boosted to a 350V DC bus voltage by push-pull boost and full-bridge rectification after filtering, and then converted into 220V/50Hz power frequency AC through a full-bridge inverter circuit. The sampling circuit samples the corresponding points to achieve closed-loop control and protection functions.

Under high current conditions, a small duty cycle of the power tube drive signal will cause severe heating and reduced efficiency. Therefore, the front stage of the inverter power supply adopts a quasi-open-loop control method, that is, when the input voltage is within a certain range, the drive signal duty cycle is opened to the maximum and remains unchanged. When the input voltage is too high, the duty cycle is reduced to maintain the bus voltage within a certain range. The advantage of this is that the front stage boost can achieve higher efficiency. System hardware design

The hardware structure of the inverter power supply is shown in Figure 2. It mainly includes a DC push-pull boost circuit, a sinusoidal inverter circuit, an output filter circuit, a drive circuit, a sampling circuit, a main controller and a dot matrix liquid crystal. Among them, the DC boost part increases the input voltage to a bus DC voltage above the peak value of the output sinusoidal AC power, the sinusoidal inverter part inverts the bus DC voltage and obtains a sinusoidal AC power through the output filter circuit, and the sampling circuit samples the bus voltage, bus current, output voltage, output current, and input voltage to achieve short-circuit protection, overvoltage and undervoltage protection, overcurrent protection, closed-loop voltage regulation and other functions. The function of the drive circuit is to match and amplify the logic level of the drive signal to meet the requirements of the drive power tube. The function of the control circuit is to generate a drive signal and process the sampled signal to achieve complex system functions. The function of the dot matrix liquid crystal is to display system working information, such as output voltage, current and protection information.

1) Main Controller

The main controller uses the STM32F103VE enhanced microcontroller. The STM32 series microcontroller is a product designed by STMicroelectronics for high-performance, low-cost, and low-power embedded applications. This microcontroller adopts the Harvard structure, which allows the processor to perform addressing and data reading and writing operations at the same time, and the processor performance is as high as 1.25 MIPS/MHz. It supports single-cycle hardware multiplication and division, with a maximum clock frequency of 72 M, and a maximum of 512 kB on-chip Flash and 64 kB on-chip RAM. It also has many peripherals suitable for inverter and motor drive, such as up to 30 PWMs and 3 12-bit precision ADCs. In this system, it is used to generate PWM and SPWM drive signals, and process the sampled signals to complete the voltage stabilization feedback and protection functions, and drive the dot matrix LCD to display system information. Considering the actual speed of the power tube and the driver chip, the frequency of the boost PWM wave is 20 kHz, and the frequency of the inverter SPWM wave is 18 kHz. According to different modulation methods, the SPWM drive signal form can be divided into: bipolar, unipolar, and unipolar frequency multiplication. Since bipolar modulation has low distortion, SPWM adopts bipolar drive mode in this design.

2) Dot matrix LCD

The LPH7366 dot matrix LCD is used, which has the characteristics of ultra-low power consumption. It is used to display the current working status of the system, such as output voltage, output current, input voltage and other information. At the same time, it indicates whether the system is in protection and what protection status it is in.

3) Auxiliary power supply

Different voltage requirements are provided for different parts of the system. The DC input voltage is stepped down to 5.0V by LM2596-5V, and then part of it is used to power the sampling circuit. The other part is stabilized to 3.3V by LDO regulator LM117-3.3V for the processor and dot matrix LCD. At the same time, an auxiliary winding of the push-pull transformer obtains a voltage of about 20V, which is rectified and filtered and stabilized to 15V by LM2596-ADJ for use by the drive circuit.

4) Driving circuit

The high-speed optocoupler isolation IGBT/MOSFET driver chip TLP250 produced by Toshiba Semiconductor is selected. TLP250 has the characteristics of high isolation voltage, strong driving capability, and fast switching speed. The schematic diagram of the drive circuit is shown in Figure 3.

In the push-pull boost driver (U1, U2), TLP250 is responsible for matching the drive signal amplitude with the current. For the full-bridge inverter driver (U3, U4, U5, U6), not only the drive level and drive capability must be considered, but also the isolation of the upper and lower tube drive signals. To simplify the design, the upper tube driver (U3, U5) of the full-bridge inverter uses a bootstrap power supply to reduce the number of isolated power supplies used.

For the inverter bridge drive circuit, in order to avoid direct conduction between the upper and lower tubes, the dead zone problem needs to be considered in the design. The PWM module of the STM32 microcontroller has a dead zone function, and this design adopts the software dead zone method. Another advantage of this is that for different power tubes, only the software design needs to be changed to obtain the best dead zone parameters.

5) Sampling circuit

Output voltage sampling is used for feedback voltage regulation, output current sampling is used for overload protection, bus current sampling is used for short circuit protection, bus voltage sampling is used to limit the false high bus voltage, and input voltage sampling is used for input overvoltage/undervoltage protection. Current transformers and voltage transformers are used in output sampling, which greatly reduces system interference and improves system reliability. The schematic diagram of the sampling circuit is shown in Figure 4.

For output current sampling, a 5 A/5 mA current transformer is used in this design. Since the output of the current transformer is a millivolt AC signal, in order to be collected by the AD module inside the microcontroller, it must be rectified into a DC signal and amplified. The ordinary diode rectifier circuit is ineffective for millivolt voltages, so a small voltage rectifier circuit composed of an operational amplifier (U11, LM3 58) is used here. Actual tests show that this circuit effectively solves the sampling problem of millivolt signals. System software design

In order to improve the readability and code efficiency of the system, the software is designed with a state machine concept. Figure 5 shows the state transition diagram of the system. After the system is powered on and reset, it enters the SAMPLE sampling state. If the sampling completion flag FINISH is detected, it enters the JUDGE state for judgment. If FAULT is not 0, that is, there is a fault signal (overvoltage/undervoltage, overload, short circuit), it enters the PROTECT state to turn off the output, and jumps to the WAIT state to wait for the fault signal to be eliminated. When the fault signal is eliminated, the system soft restarts and starts new sampling and detection. If no fault signal is detected after the JUDGE state, it enters the NORMAL normal state and adjusts the voltage.

After the system is powered on, the initialization of each peripheral is completed first, mainly including the initialization of the system clock, timer, GPIO port, ADC, DMA, interrupt and SPI. Here, once the timer and interrupt are initialized, PWM and SPWM waves will be generated. Considering the high real-time requirements of overcurrent, short-circuit protection and feedback voltage regulation, they are completed in the interrupt. Undervoltage and overvoltage have low real-time requirements and are placed in the main program. In order to improve the performance of the system, ADC sampling uses DMA to transmit data. After the transmission is completed, an interrupt request is issued to perform simple filtering on the collected data. Other function functions call this data to complete the corresponding protection and voltage regulation functions. The flow chart of the main program is shown in Figure 6.

Debugging and Experimentation

According to the above ideas, a 400 W prototype was trial-produced, using IRF3205 as the push-pull boost power tube, HER307 as the rectifier diode, and IRF840 as the full-bridge inverter power tube. The frequency of the PWM wave of the front stage boost is set to 20kHz, the frequency of the SPWM wave of the rear stage is set to 18kHz, the output filter inductor L is 1 mH, and the output filter capacitor C is 4.7μF. The actual test of the sinusoidal AC output voltage accuracy is 220 V±1%, the frequency accuracy is 50 Hz±0.1%, the THD is less than 1.5%, and the inverter efficiency is greater than 87%. The test waveform at full load is shown in Figure 7 (the output is measured by a 20 kΩ/100 kΩ resistor divider).

Conclusion

This paper discusses the design of a digitally controlled sine wave inverter power supply with STM32 microcontroller as the main controller, and discusses the key issues involved in detail. In view of the higher requirements of high-end electronic equipment for inverter power supplies, an effective solution is proposed. The use of this design scheme not only simplifies the hardware design of the inverter power supply , but also greatly improves the quality and performance of the power supply , and has a high promotion value.