Sensorless AC Motor Control


Introduction

World-wide interest in water and energy conservation as well as the overall environmental friendliness of new products and services definitely impacts the home appliance market. The actual regulatory standards and recommendations, either nationwide or multination (EU standards), enforce basic demands for a new generation of home appliances, particularly washing machines, dryers, dish washers and refrigerators. To meet these demands while still reducing system costs, enhanced microcontrollers equipped with appropriate software must be employed.

For the home appliance industry, a digital signal controller (DSC) based solution is proposed, combining the processing power of a digital signal processor with the functionality and ease of use of a microcontroller on a single chip. A flexible set of peripherals allows the designers to realize numerous functions, such as standard motor drive algorithms, advanced control algorithms, sophisticated feedback signal sensing, power factor correction schemes and communication with external environments.

Employing variable speed motor drives gives the designer an opportunity to use more sophisticated control programs. This enhances appliance performance and increases overall energy efficiency. Thus, home appliances equipped with variable speed drives and intelligent controls outperform those with uncontrolled, fixed-speed motor drives.

The majority of electric motor drives in home appliances are controlled very simply by either fixing the motor speed at pre-defined levels or by running it directly from the AC main supply without additional control electronics. The alternating current mains-supplied single-phase induction motor (ACIM) is extensively used because it is low cost, robust and reliable. However, ACIM solutions also have significant drawbacks, such as lower efficiency and reduced effective speed control, which hinder control improvements designed to enhance product flexibility and meet regulatory guidelines as well as customer expectations. In contrast to single-phase ACIM, the variable speed drives address energy efficiency requirements by maintaining precise torque control, thereby improving overall appliance efficiency.


AC Motor Choice

There are different motor categories that can drive home appliances, such as ACIM, permanent magnet (PM) motors or switched reluctance (SR) motors. The product of the electronic drive efficiency and the electric motor efficiency determines the overall energy efficiency of the system.

In general, to achieve variable speed with an ACIM, variable frequency and voltage need to be supplied to the ACIM. This is referred to as constant volt per Hz control. The ACIM speed drive efficiency can be further improved by using vector control. This requires accurate velocity information sensed by a speed or position sensor attached to the rotor. However, the additional sensor, connector and associated wiring increase the cost of the motor drive.

PM motors use permanent magnets to establish the flux instead of creating it from the stator winding. Replacing the electromagnetic excitation with permanent magnets has several advantages. Most obvious is the absence of excitation losses, which means PM motors have higher power density than comparable DC motors. PM motors have no mechanical commutator, and power density exceeds the AC induction motor because there is no flux oriented current. Overall efficiency approaches 90 percent while single-phase ACIM efficiency reaches only about 70 percent. Permanently excited synchronous motors are very attractive solutions for home appliances, however, they cannot be operated by applying AC main supply to the stator winding.


Variable Speed Motor Drives

Experts say that the total electric power consumption can be reduced up to 30 percent if electric motors are controlled to optimize the supplied energy. The controller for a variable-speed drive actually controls the rotational speed of an AC electric motor by regulating the frequency of the electrical power supplied to the motor. Electronic variable speed drives allow electric motors to continuously operate over their full speed range.

As Figure 1 indicates, the complete system for a variable speed drive includes the EMI filter, the input rectifier, the on-board DC power supply, a DSC, the signal conditioning circuits, the power inverter and gate driver.



Motor Control Strategy

The open loop scalar control represents the most popular control strategy for variable speed drives predominantly used with ACIM. As depicted in Figure 2, the scalar control is based on the variation of supply voltage frequency. Voltage magnitude is proportional to the voltage frequency and changes according to the frequency variation. This technique requires modest computational power that can be handled by an 8-bit microcontroller.


The great advantage of this simple method is that it is a sensorless mode, and the control algorithm does not need information about the angular speed or actual rotor position. Speed dependence on the external load torque, however, is a big disadvantage, which results in reduced dynamic performance. Because of this, electric motors under scalar control must be oversized in order to deliver the required torque during load transients. In addition, scalar techniques result in an inefficient system, power factor degradation in the utility's network and noisy operation. In this control approach the energy efficiency might degrade down to the 50 percent range from theoretical maximum. Motor model-based method estimates an electromotive force in which the electrical position information of machine is encoded. The estimator depicted in Figure 3 consists of the stator current observer with knowledge of RL motor parameters, fed by applied motor voltage and measured current.

Yet, market demands require the highest possible dynamic performance and operating speed range. Vector control (field oriented control) of AC machines, as a novel approach in electrical drives, provides very good performance in comparison to scalar control. Vector control eliminates most of the disadvantages of constant volt-per-hertz control.

Within a vector control system the synchronous frame current regulators have become the industry standard for inverter current regulation. The inverse Park (aß/dq) reference-frame transformation function calculates the stator currents and voltages in a reference frame synchronized to the rotating rotor field. All the electrical variables have DC steady state values when viewed in a rotating reference frame, enabling a simple PI regulator to provide zero steady state error. Additionally, it is possible to set up the coordinate system to decompose the current vector into a magnetic field generator and a torque generator. The structure of the motor controller is shown in Figure 3.


The inner current loop calculates the direct and quadrature stator voltages required to create the desired torque and flux currents. The Park (dq/aß) functions transform these voltages into three-phase AC stator voltage demands in the stationary reference frame. The motor currents are sinusoidal, thus producing smooth torque, which minimizes acoustic noise and mechanical vibration. The outer velocity loop adjusts the applied torque magnitude, which is directly proportional to quadrature torque-current and enables maintaining the required angular velocity. In order to effectively extend the operating speed range above base speed, an additional flux-weakening loop is added, which manipulates directional stator flux-current.


Sensorless Control

To run an AC motor in vector control mode, it is important to synchronize the frequency of the applied voltage with the position of the rotor flux coming from permanent-magnet in the rotor. This results in a sensorless mode of operation where the speed and position calculation algorithm replaces the sensor. The sensorless control approach to AC PM motors is an innovative motor drive feature that improves reliability and maintains high performance levels without additional cost. The system should be more accurate and efficient, smaller, lighter, less noisy and should have more advanced functionality at a lower cost. Sensorless algorithms can be broadly divided into two major groups-those that utilize magnetic rotor saliency for tracking rotor position and those that estimate rotor position from calculated motor models.

In a refrigerator, conventional control turns the compressor on and off to maintain the temperature in a predefined range. With sensorless PM motor control, the controller can accelerate the compressor to its target speed within a few seconds and can regulate speed to within 1 percent of its target. The smooth running of the compressor reduces audible noise, and the lower operating speed helps minimize the temperature cycles in the refrigeration compartment, improving overall efficiency.

Normally, a washing machine operates in two basic cycles- tumble-wash and spin-dry. During the tumble-wash cycle, the motor drive operates at low speed with high torque. During the spin-dry cycle, the motor drive works at high speed for short time periods. In newer models, the estimated speed ripples and calculated load torque provide valuable information on the washing load distribution. The speed ripple is used to estimate the load imbalance before starting the spin cycle. A variable speed motor can be used to regulate the clothes tub rotational speed and direction to redistribute the clothes to correct the imbalance.


MC56F80xx Family of DSCs

Freescale's 56800E core is an ideal solution for this particular application. It processes all motor control functions, including space vector modulation, vector control, current, flux and velocity loop control. Digital control improves drive reliability by reducing the number of discrete components found in early designs and facilitates advanced algorithms for optimal motor performance. Sensorless speed vector control discussed above is implemented on the MC56F8025 DSC with all control routines using C-callable, optimal assembler language with fractional numerical representations.


The MC56F80xx family members provide these peripheral blocks:

  • One pulse width modulation (PWM) module with PWM outputs, fault inputs and fault-tolerant design with dead-time insertion, supporting both center-aligned and edge-aligned modes
  • 12-bit analog-to-digital converters (ADCs), supporting two simultaneous conversions. ADC and PWM modules can be synchronized
  • One dedicated 16-bit general-purpose quad timer module
  • One serial peripheral interface (SPI)
  • One serial communications interface (SCI) with LIN slave functions
  • One inter-integrated circuit (I2C) port
  • On-board 3.3 V to 2.5 V voltage regulator for powering internal logic and memories
  • Integrated power-on reset and low-voltage interrupt module
  • All pins multiplexed with general-purpose input/output (GPIO) pins
  • Computer operating properly (COP) watchdog timer
  • External reset input pin for hardware reset
  • JTAG/On-Chip Emulation (OnCE™) module for unobtrusive, processor speed-independent debugging
  • Phase-locked loop (PLL) based frequency synthesizer for the hybrid controller core clock, with on-chip relaxation oscillator

The application uses the ADC periphery to digitize analog feedback signals (voltage, current) before being processed. The ADC trigger is synchronized to the PWM reload flag, but the sampling time instances vary as a function of the actual PWM pattern. This configuration allows multiple conversions of the required analog values for the DC-bus current and voltage within the one PWM cycle.

The PWM module is capable of generating asymmetric PWM duty cycles in the center-aligned configuration. This feature enables three-phase current reconstruction in critical switching patterns.

The quad timer is an extremely flexible module, providing all required services relating to time events. The application requires two of the quad timer's channels:

  • One channel for PWM-to-ADC synchronization
  • One channel for system velocity control loop (1 ms period)

The application allows communicating with the front-panel master control over an isolated serial link. This allows speed profile information to be downloaded to the controller and motor speed and torque information to be uploaded to the home appliance's master controller.


Conclusion

It is widely acknowledged that home appliance demands drive technological advances and, in turn, the production of next-generation products. In response to this trend, vector control with sensorless functionality is a technique used for this purpose, and since it eliminates mechanical sensors, it enables a cost-effective solution. Even such demanding control techniques can be realized using the low-cost DSC based on the Freescale 56800E core, which combines both MCU and DSP capabilities.

The results also indicate satisfactory performance for other various motor drive applications, such as pumps, fans, blowers, dryers, etc. MC56F80xx devices designed with a focus on motor control applications are being integrated into many home appliance product designs, particularly where national or multinational regulations on power consumption are strict.

Freescale offers a comprehensive portfolio of embedded chips, application software and development tools that are optimized for electronic motor systems. These embedded solutions are designed to allow customers to extract a high level of functionality and performance out of highly reliable, cost-efficient motors. A digitally controlled motor in a washing machine is designed to provide a more efficient agitation cycle so that less water is needed and may allow a short-high speed spin cycle, resulting in drier clothes and less energy consumption. The chips incorporate motor-specific functions which are designed to enable the appliances and other motor-driven products to operate more efficiently and reliably.


About the Author

Peter Balazovic is a system application engineer with the motor control and power efficiency group for the Consumer and Industrial markets. He works closely with the DSC/MCU product line for motor control applications and designs motor control algorithms at the Roznov Czech System Center. He joined Freescale Semiconductor in 2001.


Return to Top Return to Top