Interface Tradeoffs Optimize Motor Drive Power

Dec. 1, 2001
Interfacing a computer/controller to a power electronics driver stage.

While every motion system design is unique, the use of a microcontroller unit (MCU) or digital signal controller (DSC) to execute motion algorithms is becoming more common. Dramatically decreased costs for these controllers now provide the opportunity to integrate their benefits into many low-cost applications.

Understanding the techniques used to interface the logic-level I/O of an MCU or DSC to a power electronics driver stage is important knowledge for the system designer. With this type of system interface design, you can make interface component selection decisions based on answers concerning the type of motor, the algorithm used to control the motor, interface requirements that can be reduced by the controller peripherals, electrical safety standards, and the design for product development.

Gate Drive Interface

You can use the traditional half-bridge output topology shown in Fig. 1 to drive many motor types, including brushed-dc, brushless dc (BLDC), switched reluctance (SR), ac induction, and permanent magnet ac motors. The power stage requires a gate drive interface that, at a minimum, provides:

  • Level-shifted logic-level outputs of the MCU generating 10V to 15V between the gate and source of the output transistors.
  • Sufficient drive current during turn on and turn off to overcome Miller capacitance effects.

The high-side output device always presents a problem for the gate drive interface. It's desirable to use N-channel devices for the high-side and low-side output devices in a power stage. For a given die size and breakdown voltage, a P-channel device has a higher on resistance than the equivalent N-channel device. Using a P-channel device simplifies the gate drive circuit, but increases design costs.

Generating a gate supply voltage for the low-side devices in the power stage is easy, since the low-side devices have a circuit ground reference. However, you must reference the high-side transistor gate control voltage to the source voltage, which swings from rail-to-rail. Therefore, the high-side devices in the power stage require a gate voltage supply that can float on the source voltage.

Today, a selection of inexpensive, off-the-shelf ICs is available to simplify the design job of the gate driver circuit. Some of these ICs are relatively simple high current drivers without the level shifting circuitry required for the high-side device. Other ICs include the level shifting circuitry used to directly interface the logic and power devices. The type of gate driver you select may depend on the isolation requirements of the design. Opto-couplers electrically isolate the logic level circuits from the ac line potential — satisfying the level shifting requirements and allowing the use of simpler gate driver ICs.

An ac line-operated, full-wave rectifier with output filter provides the power bus for many motor drive applications. The low side of the rectifier becomes the 0V reference for the entire application. However, this 0V reference is not at a ground potential. Instead, there's an ac voltage on the low side that fluctuates between 0V and the peak line voltage. In many low-cost applications, it makes sense to simply float the MCU or DSC on this low-side potential. However, it's wise to add the signal isolation for safety reasons if the design will require testing or field service. At a minimum, the motor drive hardware used during the product development cycle should have signal isolation. Otherwise, the engineer that hooks up an oscilloscope to a non-isolated circuit is in for a nasty surprise.

The use of isolation circuits is applicable for “damage containment” reasons. Even if the feedback signals aren't isolated in a particular design, it may make sense to isolate the gate control signals. For example, the power devices could fail shorted, coupling the dc bus voltage back through the driver circuits and into the logic level devices — possibly wiping out a p. c. board worth of components.

Gate driver ICs are available with a variety of additional features, including undervoltage lockout, dead-time insertion, cross-conduction protection, and automatic current shutdown. The controller chosen for the design may integrate some or all of these functions into the peripherals.

You can produce the gate drive power supply by a number of methods. Ultimately, the high-side driver circuit must generate a voltage that is 10V to 15V higher than the dc bus voltage for the output stage. A bootstrapped power supply is probably one of the cheapest ways to generate the gate drive for the high-side device because a floating power supply isn't required. The bootstrap circuit charges a capacitor floating on the source voltage of the high-side output transistor (Fig. 2). Only by turning on the bottom transistor does the bootstrap circuit charge — pulling the source of the high-side transistor to 0V. The capacitor must store enough charge to maintain the required gate drive voltage during the upper transistor's on time. Consider the performance tradeoff with the bootstrap power supply because the top transistor cannot remain on for an indefinite period. Eventually the bootstrap capacitor voltage decays, and the high-side device turns off.

The type of motor employed determines the side effects of a bootstrap power supply. For motors driven with sinusoidal currents, the bootstrapped power supply may limit the range of PWM duty cycles applied to the inverter. If required, change the sizing of the bootstrap circuit components to increase the available range of duty cycles. However, the commutation requirements of BLDC and SR motors generally don't permit use of a bootstrap supply.

If the gate drive for the high-side device must be continuous, implement a floating power supply that can generate a voltage 10V to 15V above the dc bus voltage. One solution implements a charge pump circuit referenced to the source of the high-side transistor. Another technique modulates the gate drive signal with a high-frequency signal present whenever the gate drive signal is active. Fig. 3 shows the modulating signal coupled to the gate and source of the transistor via a transformer. Here, the signal rectifies on the secondary side, producing the gate drive voltage. Both techniques increase design costs.

Motor Feedback Signals

Motor and power electronics circuits require a variety of feedback signals, depending on the type of motor and the selected control algorithm. For a given feedback signal, there are usually multiple methods to acquire the signal. Many motor control algorithms need to know the phase currents in the load. Hall Effect current transducers are the easiest way to measure the phase currents. These sensors are inherently isolated from the high voltage circuitry driving the motor, which operate on the same voltage source powering the logic circuitry, and require minimal components to interface to the A/D converter on the MCU or DSC. Here, cost is the downside.

You can also measure the phase currents using PWM current sensor ICs that measure the voltage drop across a current sense resistor in series with the load. These devices float on the rail-to-rail voltage swing of the power stage output and operate from a bootstrapped power supply. The output from this current sensor is a PWM signal with a duty cycle proportional to the current in the sense resistor. The output may also connect to the controller. One approach filters the PWM output using an RC filter network that converts the signal back to an analog voltage. The downside is the control algorithm may not be able to tolerate any ripple or phase errors from the filter network. The filter components also increase design costs.

An alternate interface solution connects the PWM output of the current sensor directly to an input capture peripheral on the controller, as shown in Fig. 4, on page 16. The input capture peripheral captures the count value of a free running digital timebase on a rising edge of the input signal, falling edge, or both. You can then process the captured count values in the application software to determine the input signal period, frequency, or duty cycle.

Digital conversion provides a good alternative to analog isolation amplifiers in a design with available input capture pins on the controller device. First, convert the analog signal to the digital domain using V/F or voltage-to-PWM converters. Then, using digital opto-couplers, pass the signals across the isolation barrier. The combined V/F converter and opto-isolator solution may cost less than the analog isolation solution.

A third way to measure the phase currents uses sense resistors installed in series with the source of each low side transistor in the power stage, as shown in Fig. 5. A differential amplifier circuit that's connected to an A/D input amplifies the voltage across the resistor. When using this current measurement technique, you must synchronize the A/D conversion to the PWM signal that controls the transistor. To ensure an accurate current signal, make this measurement with the low-side transistor in the output stage turned on. A controller device with built-in A/D synchronization logic is helpful when using the shunt current measurement technique.

Software Development

The development of motion system software with traditional emulation devices is always tricky. An emulator allows the user to halt the application software at any time to look at register values, code execution history, etc. Unfortunately, stopping the software at an arbitrary point is harmful to the motor and power electronics. PWM control values will no longer update when halted — resulting in high dc currents in the motor and power stages. To combat this problem, it's best if the emulator puts the PWM signals in a state that will not cause damage to the load. For example, the dsPIC30F family PWM peripheral can optionally place all PWM output pins in the inactive state when halting the emulator. This action will turn off all output devices and allow the motor to coast to a stop. During the product development cycle, it may be useful to incorporate extra hardware features to protect the power stage. These features protect the hardware from software mistakes during product development. You could remove these features in the production version of the design to lower the hardware cost. Examples include hardware current limiting, bus overvoltage protection, and cross-conduction lockout.

The choice of controller will impact the selection of components connecting it to the power electronics stages. A general purpose MCU may fit the algorithm processing requirements, but it may not have peripherals directly meeting the needs of a motion system application. Additional hardware may be required in the interface circuit to protect output devices or provide additional signal conditioning for feedback signals. Devices targeted at motion system applications, such as the Microchip dsPIC30F 16-bit digital signal controller family, have specialized motion system peripherals that significantly reduce the complexity of the power stage interface.

Note: The Microchip name and logo, PIC, and PICmicro are registered trademarks of Microchip Technology Inc. in the USA and other countries. dsPIC is a trademark of Microchip Technology Inc. in the USA and other countries. All other trademarks are the property of their respective owners.

For more information on this article, CIRCLE 332 on Reader Service Card

Sponsored Recommendations

Comments

To join the conversation, and become an exclusive member of Electronic Design, create an account today!