Design of high-precision servo position loop based on TMS320F2812

introduction

Machine tools are the mother machine and the engine of the equipment manufacturing industry. my country's "Eleventh Five-Year Plan" development plan clearly stipulates that the domestic market share of domestic CNC machine tools must reach 60%, and the gap between high-end products and international advanced levels must be narrowed to within 5 years.

As an important functional component of CNC machine tools, permanent magnet synchronous motors Servo drive devices are one of the key basic technologies for CNC machine tools to move towards high speed, high precision and high efficiency. With the application of new microprocessors, power electronics technology and sensor technology in servo drive devices, the performance of servo drives has been greatly improved. For example, Japan's Yaskawa Company uses new microprocessors and expands new control algorithms to increase the speed frequency response to 1.6kHz. It has the function of automatically measuring mechanical characteristics and setting the required servo gain, realizing the "online automatic adjustment function"; Fanuc's new generation of drives uses a high-resolution encoder of 16 million per revolution and high-precision current detection to achieve high-speed and high-precision servo HRV (high response vector) control algorithm, reducing the maximum control current of the servo motor by 50% and reducing the motor heating by 17%, so that the servo drive device can obtain higher rigidity and overload capacity. In terms of high-performance servo drive technology, there is still a large gap between China and foreign famous brands, which has become a "bottleneck" problem restricting the development of China's high-end CNC system industry.

In view of the shortcomings of the old products, such as long signal processing time and low current and position signal detection accuracy, this system uses TMS320F2812 DSP as the controller, shortens the signal processing time and improves the current sampling accuracy; the position detection uses Tamagawa's TS5667N120 17-bit absolute encoder to improve the position detection accuracy. In the application of CNC machining centers, the system has positioning without overshoot, high rigidity, and high speed stability, reaching the design indicators and meeting the requirements of micron-level machining accuracy.

System hardware design

The system hardware uses TMS320F2812 DSP controller, Mitsubishi IPM power module, and Tamagawa TS5667N120 17-bit absolute encoder as the main functional components. The hardware system block diagram is shown in Figure 1.

In Figure 1, the TMS320F2812 DSP is the control core, which receives information from the CNC, encoder interface, current detection module and fault signal processing module to complete the control and fault processing of the permanent magnet synchronous motor . The photoelectric isolation module serves as the interface between the electronic circuit and the power main circuit, and sends the SVPWM signal sent by the DSP to the IPM module to complete the DC/AC inversion and drive the motor to rotate. The encoder interface sends the information such as the magnetic pole position, motor direction and encoder alarm of the permanent magnet synchronous motor recorded by the absolute encoder to the DSP , and sends the position information of the permanent magnet synchronous motor to the CNC. The motor phase current is measured, filtered, amplitude converted, zero offset, and limited by the current detection module, and converted into a 0~3V voltage signal and sent to the A/D pin of the DSP . The overvoltage, undervoltage, short circuit, power failure and IPM fault signals of the power main circuit are detected and processed by the fault detection module and sent to the I/O port of the DSP . The keyboard and display module is the human-machine interface of the controller, which is used to complete the input of control parameters and display the operating status and operating parameters. The memory module is used to store control parameters and system fault information.

System software design

According to the task division, the system software consists of tasks and task management modules. The task management module schedules and manages four tasks, including human-machine interface, control algorithm, acceleration and deceleration control, and fault handling. The control algorithm mainly includes: regulator control algorithm, vector control algorithm and digital filter algorithm.

According to the structured programming method and the principle of "functional independence", the system software is divided into two parts: the main program module and the vector control program module. Each part is divided into several sub-modules to facilitate software design, debugging, modification and maintenance. The vector control software design adopts a typical foreground and background mode, with the main program as the background task and the interrupt service program as the foreground task. According to the characteristics of the vector control algorithm, the interrupt service program only processes the PWM control subroutine with high real-time performance, and puts a series of low-real-time tasks such as some system measurements, keyboard processing and display into the background tasks.

The main program is the main framework of the software. Its working process is: after the system is powered on and reset, it will initialize the on-chip peripherals, read the control parameters from the E2PROM, and display the initial information on the LED. After the initialization is completed, the main program will loop to execute LED display, keyboard processing, and parameter calculation and storage.

PWM interrupt service. When the PWM interrupt arrives, the coded signal is read first, the angle and speed are calculated, then A/D sampling is performed and clark and park transformations are performed, then PI adjustment and inverse park transformation are performed, and finally the space vector module is entered to generate a PWM signal.

Controller Algorithm

The system adopts a three-loop control structure, the current loop and speed loop adopt PI control, and the position loop adopts proportional plus feedforward compensation control.

PID Control Algorithm

PID control algorithm is the most commonly used algorithm in control. For most control objects, PID control can achieve satisfactory results. In order to prevent the PID regulator from oversaturation, the system uses a PID controller with desaturation, as shown in Figure 2.

The discrete PID control algorithm is as follows:

Where, is the output before saturation, KP is the proportional gain of PID control, Ti is the integral time constant of PID control, Td is the differential time constant of PID control, and Kc is the desaturation time constant.

Control Algorithm of Position Controller

The position controller adopts a proportional plus feedforward control structure, as shown in Figure 3, where Gm is the transfer function of the motor, Gspd is the transfer function of the speed loop, Gpos is the transfer function of the position loop, and Fpos is the position feedforward controller transfer function.

The transfer function of the system is:

When Fpos(s)=1/(Gspd(s)Gm(s)), H(s) =1, the output can completely reproduce the input signal, and the transient and steady-state errors of the system are both zero. When the speed regulator adopts PI control, when the cutoff frequency of the position loop is much smaller than the cutoff frequency of the speed loop, the speed loop can be equivalent to an inertia link, and the motor can be equivalent to an integral link, so Fpos(s) can be regarded as two parts: acceleration feedforward and speed feedforward [5], where: The differential equation of the acceleration term in the position feedforward is:

Where R(k) is the position given signal in the Kth sampling period; Yaf is the output of the acceleration signal in the Kth sampling period, and Kaf is the acceleration feedforward proportional coefficient.

The velocity term difference equation in position feedforward:

Where R(k) is the position given signal in the Kth sampling period; Yaf is the output of the speed signal in the Kth sampling period, and Ksf is the speed feedforward proportional coefficient.

The corresponding difference equation of the position loop P is:

Where R(k) is the position given signal in the Kth sampling period; C(k) is the position feedback signal in the Kth sampling period, Ye is the output of the position loop signal in the Kth sampling period, and Kc is the position loop proportional coefficient.

Absolute encoder communication program

Absolute encoder The interface with DSP uses CPLD as the interface chip. The CPLD program is written in VHDL language, and the program structure is shown in Figure 4. This circuit completes the conversion of serial input data to parallel output data, and the conversion of parallel input data to serial output data.

In Figure 4, module DIV is a clock divider, and the TX module receives 8-bit parallel data from the microprocessor interface module MP and outputs the data serially to the RS-485 port through port DOUT. In turn, the RX module receives serial data input and sends it to the MP module in 8-bit parallel format. The MP module also converts the received position signal into a pulse form output to achieve connection with the CNC.

Experimental results analysis

In this design, virtual instrument technology is used to design an experimental test platform and record the experimental test results. The virtual test platform is configured as follows: software NI LabVIEW 8.0, hardware NI M series multi-function data acquisition card PCI-6251, 16, NI counter/timer PCI-6602.

Figure 5 shows the speed waveform during the processing. Figure 5 shows that the acceleration and deceleration time of the system is less than 200ms; there is no position overshoot; when stable, the speed fluctuation is less than 0.1 revolution. Speed ​​frequency response: greater than 300Hz; speed fluctuation rate: less than ±0.01% (load 0~100%), 0 (power supply ±10%); speed regulation range: 0.1rpm~3000rpm; rotation positioning accuracy: 1 pulse.

Figure 6 shows the machining results of the drive with a domestic brand machining center. Experimental test data: the surface roughness of the upper surface is Ra1.6μm; the roughness of the side surface (i.e. the measuring surface) is Ra3.2μm.

Conclusion

For the feed control of CNC machine tools, magnetic field oriented control and feedforward compensation control are adopted. With TMS320F2812 DSP controller, IPM power module and TS5667N120 17-bit absolute encoder as the main functional components, the designed permanent magnet synchronous motor servo drive controller has the advantages of positioning without overshoot, high rigidity and high speed stability in the application of CNC machining center, which meets the design indicators and can meet the requirements of micron-level machining accuracy.