Integrate motor control and power factor correction with DSP controller

    As digital signal processor (DSP) prices drop from several hundred dollars to three dollars, DSPs are increasingly being adopted in price-sensitive home appliances such as washing machines, refrigerators, heaters, ventilation and air conditioners. High MIPs DSPs with special peripherals, in addition to significantly improving the performance of these products, also greatly simplify the product design process and provide a variety of important special features. The extraordinary processing power of DSP enables manufacturers to meet the increasing requirements of users, such as higher efficiency and variable speed operation and precise speed control characteristics.

    Low-cost DSP controllers enable many advanced motor control algorithms to be built into very cost-sensitive applications. The bandwidth of DSP controllers also enables designers to integrate multiple functions, such as motor control, power factor correction and communications protocols, into one controller (see Figure 2). This article describes the structure of a single-board AC induction motor drive (with power factor correction and serial communication) using a low-cost DSP controller, and verifies the advantages of the DSP software and hardware solutions in detail.

Key Motor Drive Requirements

    In order to provide advanced motor drive characteristics, a typical motor controller must meet the following requirements:

    ·The ability to generate multiple high-frequency, high-resolution PWM waveforms.

    ·Can process data in real time to implement advanced algorithms to minimize torque fluctuations, online parameter adaptation, precise speed control, etc.

    ·Use the same controller to achieve multiple functions such as motor control, power factor control and communication.

    · Simplify the overall solution by reducing component count, simplifying board layout and improving manufacturing processes.

    · Simplify post-production corrections through software rather than hardware redesign.

DSP Controllers for Motor Drives

    TI's new DSP integrates peripherals for all critical power electronics to simplify overall system execution, reduce component count and reduce board size for the most cost-sensitive applications such as home appliances.

    TI's TMS320F240 DSP controller (Figure 1) provides multiple functions for a 3-phase ac induction motor drive, including basic motor control with closed-loop rate control, input power factor correction using a boost topology, and serial communications. This device has a 20MIPs 16-bit fixed-point DSP core and integrates the following power electronics peripherals: 12 PWM channels (9 independent channels), 3 16-bit multi-mode general-purpose timers, 16-channel 10-bit ADC ( With simultaneous conversion capability), 4 capture pins, encoder interface capability, SCI, SPI, watchdog, etc.

    6 PWM channels (PWM1~PWM6) control the 3-phase voltage source inverter. These 6 PWM channels are divided into 3 pairs (PWM1&2, PWM3&4, PWM5&6). 3 compare registers are associated with each PWM channel pair. Update the compare register value to get the appropriate PWM output. Programmable on-chip software deadband modules provide sufficient dead time to avoid failure shock. There are more than 3 PWM channels reserved to implement other functions, such as power factor correction.

Motor drive description

    3-phase inverter utilizing 6 PWM channels. Uses programmable on-chip software to generate deadband; 25-tooth sprocket provides speed input to controller capture unit; implements power factor correction using boost topology where another PWM channel of the controller controls the power switch . A simple voltage divider circuit is used to regulate and feed different voltages to the analog/digital unit of the controller.

    The complete driver mainly utilizes software modules: closed-loop interval vector pulse width modulation (SVPWM) for motor control; power factor correction module to improve input power factor; and serial communication.

    The implementation of the above-mentioned software modules will be described in detail below.

Spaced Vector Pulse Width Modulation (SVPWM)

    In order to provide energy efficiency to the motor, it is important to use the correct inverter switching switching method (see Figure 3). Interval vector pulse width modulation is one such switching method that is much superior to simple and inefficient methods such as sinusoidal PWM. SVPWM has higher DC bus utilization and lower harmonic copper losses. For a given dc loop voltage output, SVPWM improves the power output of its 3-phase motor by 16% compared to a sinusoidal PWM fed motor.

    All application software is driven by an interrupt service routine (ISR). The main code (i.e. local loop) consists of TMS320C240 peripheral initialization (such as phase-locked loop, watchdog, interrupt control and event management). The rest of the code consists of PWM-ISR. The ISR is called every 50ms (20KHz) by the periodic event flag in Timer1 of event management.

    The reference voltage vector needs to be accurately positioned in the dq diagram (see Figure 5). The goal is to rotate the datum vector in the dq diagram with a given speed ω and size M. The controller receives the required reference speed ω from the user side. The angular velocity ω is controlled by a precise frequency generation algorithm that relies on the modulus properties of a limited-length register (called an integrator in Figure 5). The upper 8 bits of the integrator are used as pointers to the 256-word sine lookup table. Appending a fixed value (step size) to this register causes the 8-bit pointer to cycle through the sine table at a constant rate. At the end of the limit, the pointer simply loops and continues the same operation for the next modulo value given by the step size. The Sine(a) value needs to be used to decompose the reference voltage vector into the basic interval vector in the area where the reference voltage vector is located. Since this decomposition is the same in the 6 regions, only a 60° sine lookup table is needed. In order to complete a cycle (360°), the sine table needs to be searched 6 times. For a given step size, the angular frequency of V* (cycles/second) is given by:

    ω=(STEP×fs)/6×2m (1)

    (1) where fs is the sampling frequency (20KHz, PWM frequency), STEP is the angular step increment, and m is the number of bits in the integration register.

    For a given PWM frequency (fs=20KHz), STEP (=1) and m (16 bits), the frequency resolution is 0.061Hz. Therefore, the inverter output frequency can be controlled within 0.1Hz. The size of the lookup table also has a certain impact on the harmonic distortion of the synthesized sine wave. A 256-entry table is used for the 60° range, which provides an angular lookup resolution of 60°/256=0.230.

The reference voltage vector is transformed into a set of dedicated switching variables a, b, c. The

    3-phase voltage source inverter can generate 8 basic vectors. Six of the eight basis vectors (V1, V2, V3, V4, V5 and V6) are shown in Figure 5. The other two vectors V0 and V7 (zero vectors) are not shown in the figure and correspond to states 0 (000) and 7 (111) of the switching variable. In order to generate an arbitrary basis vector V*, the given basis vectors need to be time averaged. The desired reference voltage vector V* in a given area can be synthesized as a linear combination of two adjacent basis vectors Vx and Vy, which are combined with either of the two zero vectors to form the frame area. Therefore, the reference vector can be written as:

    V*=dxVx+dYVY+dzVz (2)

    where Vz is the zero vector and dx, dy and dz are the duty cycles of states X, Y and Z within the PWM switching interval. The duty cycle must add up to 100% of the PWM period, that is, dx+dy+dz=1.

    V*=Mvmaxeja=dxVx+dYVY+dzVz (3)

    where M is the modulation coefficient.

    Decompose V* into its dq component, which can be written as:

    Solving dx and dy, we get:

    dx=Msin(60-a) (6)

    dy=Msin(a) (7)

    Since the dq reference frame can be compared to any basic vector, So these equations can be applied to any area. That's why only a 60° sine lookup table is needed in this approach.

    For a specific reference voltage V*, after calculating the required PWM duty cycles dx, dy and dz, it is necessary to calculate the appropriate comparison value of the compare register. Each PWM cycle (50ms) needs to calculate 3 new comparison values ​​(Ta, Tb and Tc) to generate the switching pattern.

    The comparison values ​​Ta, Tb and Tc can be written as:

    Ta=(T-dx-dy)/2 (8)

    Tb=dx+Ta (9)

    Tc=T-Ta (10) The

    calculated comparison value is input into the PWM comparison register, The DSP controller updates the duty cycle at the beginning of the next PWM cycle.

    Obtain motor speed information using a low-cost 25-tooth sprocket and Hall sensor. Figure 4 shows a physical description related to the sprocket and the associated angular rate. The total pulse rate is 25 pulses/cycle. The output of the Hall sensor is sent directly to the Capture1 input of the C240, where the tooth-to-tooth period (t2-t1) can be measured. To reduce jitter or cycle fluctuations, each time a new pulse is detected, the most recent 25 cycle measurements are averaged.

    Once a "robust" period measurement is derived from an averaging algorithm, the inverted period value is used to calculate the angular rate (frequency). A classic proportional-integral (PI) can be used to obtain closed-loop speed control.

Power Factor Correction

    Power factor correction is becoming increasingly important in many industries. The IEC in Europe and the IEEE in the United States have been engaged in developing standards for limiting harmonic currents in offline equipment.

    Many new electronic products must have a power factor close to 1 and a distortion-free current input waveform. A typical AC-DC converter uses a simple capacitor-based diode bridge rectifier to draw power from the ac line. The input current waveform in an offline device uses a rectifier/capacitor input filter to provide narrow pulses. Therefore, the power factor is far from ideal due to high harmonic distortion of the current waveform.

    The input power factor can be improved using the well-known boost topology with a single power switch. The power switch controls the flow of energy. When the switch turns on, current builds up in the boost inductor while the diode remains off. When the power switch is turned off, the energy present in the inductor charges the dc line capacitor through the diode. Additionally, the inductor current can be controlled to track the desired voltage. For power factor correction, the inductor current is usually controlled to track the rectified voltage, so the ac terminal current will be in phase with the ac line voltage. However, the method based on forward feed control and current measurement is not a simple implementation method. Measure the rectifier output voltage to obtain waveform information of the input line current. The full-wave rectified line voltage is fed to the analog channels. The input voltage is adjusted with a resistor divider.

    In addition, measuring the DC line capacitor voltage can obtain information about the input current. Again, use a resistor divider to feed this voltage to another analog channel.

    These two voltages determine the operating duty cycle of the boost control switch. The duty cycle can be written as:

    dactual=Krec×Vin×dmax,

    where dactual is the duty cycle of the boost power switch, dmax is the maximum allowable duty cycle, and krec is the gain of the rectifier output voltage sensor.

    The DC bus voltage level determines the maximum allowed duty cycle dmax.

    The classic PI method can obtain dmax. Vref depends on the application and is always greater than the peak value of the input line voltage. The analog/digital conversion is synchronized to the corresponding power factor corrected PWM channel. Figure 6 shows timing diagrams for different control loops.

    The size of the boost converter is much smaller than various passive filters. It can be expected that effective power factor correction will meet various stringent power supply performance adjustments in the future.

An experimental

    250W experimental board has been made to implement all the functions mentioned above. A 3-phase induction motor with fan (1/3 HP, 4 poles) is connected to the board as a load. The input is standard ac line voltage (115V, 60Hz). The operating frequency varies between 0 ~ 60Hz. IRF840 power MOSFETs are used as power switches for 3-phase inverters and boost converters. The boost inductor is around 150mH. The software is written in assembly language and uses less than 4KW program space. The controller's on-chip flash memory stores the program.

    Distortion-free phase current is generated by 20KHz SVPWM. Table 1 shows the power factor levels at different operating frequencies.

    Detects some distortion in the current waveform. The boost inductor will be in saturation state at higher operating frequencies, causing current distortion. However, the motor phase currents and line-to-line voltages are completely distortion-free.

    Soft start feature is also included in the software. The drive ramps up at a predetermined rate during startup. Make sudden changes in speed command through ramp up or down curves.

    Table 2 lists the DSP bandwidth usage for different software modules. Obviously, multi-function integration cannot utilize 100% of the DSP bandwidth, so more advanced algorithms can implement application requirements.

Table 1 Improving power factor correction using boost topology

Table 2 DSP bandwidth usage

Conclusion

    New low-cost DSP controllers are becoming a promising option for cost-sensitive applications. These controllers have sufficient bandwidth and integrate peripherals with on-chip power electronics to implement multiple functions of the motor drive. Integrated functionality reduces system component count and speeds time to market.