Theory & Practice of Electromagnetic Design of DC Motors & Actuators

George P. Gogue & Joseph J. Stupak, Jr.

G2 Consulting, Beaverton, OR 97007


CHAPTER 9

ELECTRONIC CONTROLLERS

9.1 Types of drives:

Figure 9.1 shows an elementary switching circuit for a 3-phase motor. A unipolar (half-wave) circuit is shown in (a) and a bipolar (full-wave) circuit is shown in (b).

The unipolar operation requires 3 power devices only and they are switched on and off consecutively. This means that torque is developed by one phase only at any time and the variation in torque during that period of conduction follows the torque (and emf) waveform of the particular phase carrying current.

Figure 9.1 (b) shows that for bipolar operation, two power devices e.g. Q1 and Q6 are switched on while all the other devices are off. This action connects 2 phases, i.e. A and B, in series and across the supply voltage V. The sequence of switching continues in such a way that halfway through the conduction interval of phase A, current is switched over from phase B to phase C. This continuous overlap produces less ripple in the developed torque of the motor than in unipolar operation as seen in Table 7.1. The winding in Figure 9.1(b) can easily be changed to a connected winding and still use the same switching circuit.

The control of the switching instants and the sequence of the power devices in Figure 9.1 is done with control circuitry not shown in the diagram. This circuitry consists of logic gates or a microprocessor chip. In any case, correct switching can only be achieved if accurate information is obtained about the position of the rotor with respect to the particular part of the winding being switched. This is done with position sensors being switched by the rotor and connected to the switching circuit.

9.2 Speed Control:

In most applications where a brushless motor is used, it is desirable to exercise some control over one or more of the performance parameters. Examples are current and speed control. It is necessary in either case to obtain a feedback signal from the motor corresponding to these parameters.

In the case of current control, the voltage drop across a current-sensing resistor is needed. For speed, either a voltage from a tachogenerator or a train of pulses with a frequency proportional to speed is needed. This feedback information is compared with reference values preset to the desired values and the error, if any, is used in effecting a correction to the switching circuit.

The control of the switching circuit resulting from this error signal is implemented by controlling the switching of the power devices. If transistors are used, one of two common control schemes is implemented:

a) Linear control:

In this scheme the saturation level of the transistors is controlled and the operating point on its characteristics is somewhere between the conditions of completely on or completely off. This control is done at the expense of large amounts of power dissipation in the transistor. However, there is a significant advantage in this scheme providing very smooth current waveform to the motor. The low efficiency of the drive electronics must be taken into account when calculating overall efficiency of the system.

b) PWM control:

The pulse width modulation scheme is more popular because of the high efficiency achievable. In this scheme the switching devices are switched between the on and off conditions at a much higher frequency than the basic commutation frequency. By doing so, the supply voltage is applied across the winding for short intervals as seen in Figure 9.2. This is equivalent to a mean voltage level lower than the maximum voltage and a function of the on and off time periods. The current and speed of the motor can then be controlled by controlling this effective level of voltage.

The disadvantage of the scheme is the electric noise generated in the process of switching the devices at these high frequencies. This also causes audible noise and vibration in the motor. But more importantly, the continuous switching can give rise to additional losses in the motor in the form of additional iron losses.

The eddy current and hysteresis losses in the steel as well as the eddy currents in the copper wires being dependent on frequency can see a significant rise. The copper losses can also increase as the effective resistance of the winding would increase with resistance.

These losses should be taken into account when determining the efficiency of the motor. However, it may be difficult to calculate these values accurately prior to testing.

9.3 Sensorless control:

Since their inception, brushless dc motors had to be equipped with position sensors of some kind to determine the position of the rotor. The information from these sensors is used in switching current to appropriate coils of the motor to facilitate continuous rotation and unidirectional torque. These sensors are usually supplied inside the motor with connections to the outside drive electronics. The penalties of using these sensors are many: added cost for the sensors and their support; lost space from electromagnetic and mechanical parts; additional assembly and tuning time. The above is also true if the sensors are left outside the motor and the only difference is the question of whose responsibility it is to add the sensors.

The sensorless control scheme for brushless motors eliminates the problems stated above by determining the rotor position without sensors. Information is obtained from other parameters of the motor to determine the rotor position and then to control the winding current and produce the desired torque. There are presently several methods by which position sensing without sensors is achieved and many more are still possible.

One common method is to sense the induced voltage in the winding which is not carrying current at the particular moment. The zero crossing of that voltage is electronically detected, thus establishing where the rotor is with respect to the winding and replacing the function of the position sensors.

This and other methods, however, have the common problem of not providing information at start before any motion of the rotor occurs. This is solved by electronically generating fictitious pulses until the rotor attains the desired speed at which the sensorless control scheme would take over.

The sensorless solution naturally requires more electronic circuitry than the modest amount for the standard brushless motor. It can, however, be placed conveniently with the major circuitry of the controller, away from the motor and freeing the sensor location for electromechanical parts. In addition to this improved space utilization, it is also true that time is saved in the assembly and tuning of the system due to the absence of the sensors. The reliability of the drive is improved for the same reason.

Since the sensorless control is a technique which replaces position sensing with discrete devices, it is usually contained in a separate semiconductor device, apart from the power device. This enables the user to change the switching devices for various power levels and still use the same control device.