PaperlayerView exact revision on arXiv

Source-linked AI summary

Position and Speed Control of Brushless DC Motors Using Sensorless Techniques and Application Trends

Jose-Carlos Gamazo-Real, Ernesto Vazquez-Sanchez, Jaime Gomez-Gil

arXiv:2402.05263v1eess.SY

TL;DR

BLDC drives require rotor-position information for commutation, while physical sensors introduce cost, mounting, temperature, wiring, and reliability constraints. This paper reviews sensorless back-EMF, estimator, model-based, implementation, and application techniques. Its conclusion is that sensor elimination can reduce cost and improve reliability, and is necessary in some harsh-environment applications.

  • Problem

    BLDC drives require rotor-position information, but physical sensors add cost, size, mounting, temperature, component, and wiring constraints and may be unreliable in harsh environments.

  • Method

    The paper technically reviews sensorless BLDC position and speed-control methods, emphasizing back-EMF schemes, estimators, model-based techniques, feasibility, implementation issues, and applications.

  • Results

    Sensorless control can eliminate position sensors to reduce cost and increase reliability, while serving applications where sensors cannot function reliably under harsh environmental conditions.

  • Takeaways & Limitations

    Sensorless control is a practical alternative to sensor-based BLDC drives and the only choice for some applications requiring reliable operation in harsh environments.

  • Takeaways & Limitations

    Back-EMF sensorless methods have zero signal at standstill, and specific schemes can suffer low-speed errors, sensing difficulties, or accumulated noise and offset effects.

Abstract

from arXiv · show

This paper provides a technical review of position and speed sensorless methods for controlling Brushless Direct Current (BLDC) motor drives, including the background analysis using sensors, limitations and advances. The performance and reliability of BLDC motor drivers have been improved because the conventional control and sensing techniques have been improved through sensorless technology. Then, in this paper sensorless advances are reviewed and recent developments in this area are introduced with their inherent advantages and drawbacks, including the analysis of practical implementation issues and applications. The study includes a deep overview of state-of-the-art back-EMF sensing methods, which includes Terminal Voltage Sensing, Third Harmonic Voltage Integration, Terminal Current Sensing, Back-EMF Integration and PWM strategies. Also, the most relevant techniques based on estimation and models are briefly analysed, such as Sliding-mode Observer, Extended Kalman Filter, Model Reference Adaptive System, Adaptive observers (Full-order and Pseudoreduced-order) and Artificial Neural Networks.

1. Introduction

BLDC motors gained prominence for efficient, reliable, compact drive applications, while sensorless control emerged to reduce sensor-related cost and reliability constraints. The paper reviews sensor-based fundamentals, sensorless advances, feasibility, and implementation issues.

  • 1. Introduction: Rising energy prices and expanding automotive hybrid-drive applications increased demand for efficient permanent-magnet motor drives, including BLDC motors.The introduction contrasts adjustable permanent-magnet drives with cheaper induction drives in energy-saving applications.
  • 1. Introduction: BLDC motors offer high efficiency, reliability, dynamic response, operating speed, and torque density compared with brushed DC and induction motors.Their high torque-to-size ratio suits applications where space and weight are critical.
  • 1. Introduction: Sensorless control omits position sensors and can reduce actuator cost, but requires more demanding algorithms and more complicated electronics.The paper presents this trade-off as a central motivation for reviewing sensorless techniques.
  • 1. Introduction: Position sensors add cost, size, mounting requirements, temperature sensitivity, components, and wiring, while sensor failure can destabilize control.These constraints motivated development of sensorless control for broader BLDC applications.
  • 1. Introduction: The paper covers sensor-based fundamentals, sensorless back-EMF and estimator methods, feasibility comparisons, implementation issues, and applications.The review structure includes open-loop starting and other practical control considerations.

2. Position and Speed Control of BLDC Motors Using Sensors

Conventional BLDC position and speed control uses sensors to identify rotor position and trigger six-step electronic commutation. The reviewed sensors include Hall-effect, variable-reluctance, and accelerometer devices, each with distinct integration and operating characteristics.

  • 2.1.1. Hall-effect sensors: BLDC commutation requires rotor position information because the controller energizes stator windings in sequence according to rotor orientation.Hall sensors embedded in the stator provide signals used to determine the energizing sequence.
  • 2.1.1. Hall-effect sensors: Hall sensors encode rotor position through three digital signals, whose combinations determine the six-step commutation sequence.A Hall sensor changes state every 60 electrical degrees, producing six steps per electrical cycle.
  • 2.1.1. Hall-effect sensors: Three-branch vertical Hall sensors use aligned rotor magnets and stator-mounted sensing elements to provide motion and shaft angular-position information in miniaturized motors.The arrangement requires alignment between rotor orientation, the permanent magnet, the stator, and the sensor.
  • 2.1.2. Variable reluctance (VR) wheel speed sensors: Variable-reluctance sensors detect toothed-wheel movement through magnetic-flux variation, generating pulses whose frequency and voltage are proportional to wheel velocity.They are passive, self-generating, low-cost, robust, and can operate above 300 °C.
  • 2.1.2. Variable reluctance (VR) wheel speed sensors: VR sensors are presented as the most suitable choice for rotor position and speed measurement because Hall-effect sensors are more expensive.Their small size also supports embedding in locations where other sensors may not fit.
  • 2.1.3. Accelerometers: Accelerometers measure acceleration through displacement of a proof mass, with steady-state acceleration represented by spring extension.The paper describes a damped mass-spring structure and relates spring constant, proof mass, and displacement.

3. Techniques and Advances in Sensorless Control

Sensorless BLDC control eliminates position sensors by extracting rotor-position information from electrical measurements, especially motor-terminal voltages and back-EMF. The reviewed approaches include direct and indirect back-EMF detection, with implementation based on motor models and practical sensing constraints.

  • Sensorless control rationale: Sensorless control can remove position sensors, reducing motor assembly cost and size when positioning is unnecessary and load dynamics are predictable.Back-EMF and current sensing can often estimate rotor position sufficiently for variable-speed operation.
  • Sensorless control rationale: BLDC excitation leaves one phase unexcited between commutations, allowing its back-EMF to provide a low-cost rotor-position signal.Only two of three phase windings conduct at a time, while the unused phase carries the back-EMF.
  • Motor model: Rotor position can also be inferred from position-dependent stator inductance arising from the motor’s magnetic saturation characteristics.The paper presents this relationship in the context of the sensorless BLDC drive model.
  • Motor model: The BLDC drive model relates applied voltage to winding resistance, inductance, current variation, and trapezoidal back-EMF.The analysis uses a phase-A equivalent circuit and assumes equal stator resistances, constant self-inductances, zero mutual inductance, and negligible iron losses.
  • Method taxonomy: The paper divides back-EMF sensing into direct detection and indirect detection methods.Direct methods include zero-crossing detection and PWM strategies; indirect methods include back-EMF integration, third-harmonic integration, and terminal-current sensing.

3.1. Back-EMF Zero Crossing Detection method (Terminal Voltage Sensing)

Terminal-voltage sensing detects commutation timing from the zero crossing of the floating phase’s back-EMF, commonly using a 30° delay to schedule six-step commutation. Its simplicity is offset by filtering delays, noise sensitivity, and limited low-speed operation, requiring open-loop starting.

  • Operating principle: Zero-crossing detection monitors the unexcited phase and delays commutation by 30 electrical degrees in a six-step inverter scheme.Each phase conducts for 120 electrical degrees, while commutation occurs every 60 electrical degrees.
  • Practical constraints: Low-pass filters suppress inverter-switching harmonics, but their time delay limits high-speed BLDC operation.Three filters may be used to extract the back-EMF from phase terminal voltages.
  • Voltage measurement: Terminal-voltage sensing commonly reconstructs or compares the floating-phase voltage using a virtual neutral point because the motor neutral is usually unavailable.The floating-phase voltage depends on back-EMF, neutral-point potential, and inverter transistor or diode voltage drops.
  • Operating principle: The floating-phase zero crossing occurs when its voltage reaches approximately one half of the DC rail voltage.Sampling at the end of the PWM on-state is selected because it is described as noise-free.
  • Practical constraints: The method is simple but becomes noise-sensitive and typically operates over a reduced speed range of about 1,000–6,000 rpm.Rapid acceleration or deceleration can also introduce commutation-position error, especially in low-inertia systems.
  • Practical constraints: Back-EMF zero-crossing methods cannot reliably detect position at standstill or low speed because back-EMF is proportional to motor speed.Open-loop PWM commutation is therefore used during the initial running stage, while high-speed operation may require on-time detection to allow resonant transients to settle.

3.2. Third Harmonic Voltage Integration method

Third-harmonic voltage integration extracts rotor-flux position from the summed phase voltages of a symmetrical Y-connected BLDC motor. The method needs little filtering and supports a wide speed range, but integration can accumulate low-speed noise and offset errors.

  • Signal generation: Summing the three stator phase voltages suppresses the fundamental, fifth, seventh, and other non-triplen harmonics, leaving a signal dominated by the third harmonic.The third harmonic maintains a constant phase displacement from the fundamental air-gap voltage across load and speed.
  • Signal processing: The filtered third-harmonic voltage is integrated to estimate rotor-flux linkage and determine switching instants.The resulting rotor-flux estimate supports inverter current control and commutation timing.
  • Commutation timing: The rotor-flux third-harmonic component crosses zero every 60 electrical degrees, coinciding with the desired current commutation instants.For maximum torque per ampere, stator current is maintained at 90 electrical degrees relative to rotor flux.
  • Implementation: Among resistor-network sensing options, only VSH is identified as suitable for third-harmonic sensorless BLDC operation.The alternatives use VSN, VSH, or VNH voltage measurements, and open-loop starting remains necessary for back-EMF-based operation.
  • Performance and limitations: The technique offers simplicity, low electrical-noise susceptibility, and robustness across a reported speed range of 100–6,000 rpm.Its higher third-harmonic frequency enables low-speed detection, while reduced filtering limits filtering-delay effects.
  • Performance and limitations: At low speed, integration can accumulate sensing noise and offset over long intervals, causing serious position error.An improved implementation uses third-harmonic back-EMF integration to reduce high-speed commutation retardation compared with terminal-voltage integration.

3.3. Free-wheeling Diodes Conduction Detection method (Terminal Current Sensing)

Free-wheeling diode conduction detection infers rotor position from current caused by back-EMF in the unexcited phase. It requires a dedicated starting procedure and has practical hardware limitations despite improved low-speed performance.

  • Applicability: The indirect sensing algorithms described apply only to SMPM motors whose winding inductances are nearly equal and rotor-position invariant.For IPM motors, position-dependent inductance changes the phase impedances and neutral-point potential, preventing terminal current sensing.
  • Detection principle: Rotor position is inferred from the conducting state of free-wheeling diodes when back-EMF drives current through an otherwise inactive phase.The method uses the conducting condition of antiparallel diodes connected to the power transistors.
  • Starting procedure: Position cannot be detected at standstill because the open-phase current depends on back-EMF.Starting therefore excites two arbitrary phases for a preset time before open-loop commutation advances the switching pattern by 120°.
  • Limitations: The method has transient commutation-point error and requires six isolated power supplies for comparator circuitry, limiting practical application.Despite this drawback, it outperforms earlier back-EMF methods at low speeds.

3.4. Back-EMF Integration Method

Back-EMF integration determines commutation from the accumulated signal of the silent phase rather than its instantaneous zero crossing. This reduces switching-noise sensitivity and adapts switching instants to speed, but low-speed operation remains problematic.

  • Operating principle: Commutation occurs when the integrated silent-phase back-EMF reaches a predefined threshold after integration begins at its zero crossing.The threshold corresponds to a commutation point and can be changed to provide current advance during flux weakening.
  • Operating principle: The integrated back-EMF area is approximately constant across speeds, allowing the threshold voltage to remain constant under the trapezoidal, speed-insensitive-slope assumption.The method assumes the back-EMF varies linearly from positive to negative.
  • Limitations: Low-speed operation is poor because error accumulation and offset-voltage problems affect the integration process.The approach is less sensitive to switching noise but does not eliminate low-speed limitations.
  • Implementation: The integrator resets to zero after reaching the threshold and remains disabled until residual open-phase current passes its zero crossing.This prevents the integrator from restarting prematurely.
  • Advantages: Back-EMF integration provides significantly improved performance over zero-crossing detection while reducing switching-noise sensitivity and automatically adjusting for speed changes.The rectified back-EMF feeds an integrator whose output is compared with a preset threshold.

3.5. Methods based on PWM strategies

PWM-based sensorless methods select switching and sensing strategies to recover floating-phase back-EMF while managing noise, voltage drops, speed range, and converter complexity. The reviewed approaches span neutral-point elimination, high-speed sensing, low-power operation, and reduced-switch configurations.

  • PWM control fundamentals: PWM control commutates two conducting phases while leaving one phase floating for back-EMF measurement.The inverter sequence reproduces the brush and commutator functions of a conventional DC motor.
  • Conventional sensing: Conventional terminal-voltage sensing compares the floating-phase terminal with the motor neutral point, but PWM causes neutral-point fluctuation, common-mode voltage, and high-frequency noise.Voltage dividers and low-pass filters are required to reduce these effects.
  • Technique of virtual neutral point elimination: Neutral-point elimination measures floating-phase terminal voltage during PWM off-time, when it is directly proportional to back-EMF without superimposed switching noise.The signal is ground-referenced, so neutral-point voltage information is unnecessary.
  • Technique of virtual neutral point elimination: The neutral-point elimination signal is unattenuated, unfiltered, and usable across a wider speed range, including high-speed operation.The technique is reported as applicable to both high- and low-voltage systems without voltage scaling.
  • Low-speed limitations: At low speed, device voltage drops and reduced back-EMF amplitude make zero-crossing detection uneven or difficult.Diode voltage-drop bias can create uneven zero crossings and unexpected commutation, especially during start-up.
  • Four-switch and direct-current PWM: A four-switch converter reduces the conventional six-switch configuration but has limited voltages, making 120° conduction difficult and requiring a 60° phase-shifted PWM strategy.Conventional PWM schemes cannot be directly applied to this asymmetric-voltage configuration.
  • Four-switch and direct-current PWM: Direct current-controlled PWM uses current rather than voltage control and provides robust speed and torque responses with simple hardware and software implementation.The approach is presented as a low-cost, high-performance alternative for the four-switch BLDC drive.

4. Other Sensorless Techniques: Estimation and Model-Based Methods

Estimation and model-based sensorless methods reconstruct unavailable rotor states from motor inputs and outputs, enabling position and speed control without direct sensing. The reviewed approaches include observers, EKF, MRAS, and adaptive flux observers, each balancing accuracy, robustness, computation, or speed-range performance.

  • Observer-based methods: Observers use measured system inputs and outputs to estimate otherwise unavailable motor states, including rotor position and speed.The estimated-output error is fed back into the model to correct the estimates for closed-loop control.
  • Practical limitations: Low-speed sensorless estimation remains difficult because accurate operation depends strongly on stator-resistance knowledge and parameter robustness.The resistive voltage drop becomes especially important at low speeds, whereas estimation is more accurate at higher speeds.
  • Extended Kalman Filter (EKF): Extended Kalman filters estimate rotor position and speed from measured stator voltages and currents, without filtering those signals.The approach combines motor dynamics, noise descriptions, and initial-condition information for nonlinear estimation.
  • Extended Kalman Filter (EKF): EKF-based estimates achieve sufficient accuracy in both steady-state and dynamic operation, but the method is computationally complex.The review notes that high-performance processors have reduced the practical importance of this computational burden.
  • Model Reference Adaptive System (MRAS): MRAS estimates speed and stator resistance by driving the output error between reference and adjustable motor models toward zero through an adaptation law.The review identifies fast adaptation and relatively easy implementation as key advantages.
  • Adaptive observers: The Adaptive Pseudoreduced-order Flux Observer uses Lyapunov design and a constant gain matrix to reduce computation and improve speed response over the AFFO across a wide speed range.Its gain choice keeps observer poles independent of motor speed.

5. Implementation of back-EMF Control Techniques

Back-EMF control implementation requires selecting electronic-processing hardware while addressing the need to start the motor from standstill. The reviewed implementation discussion therefore links processor or ASIC choices with open-loop starting requirements.

  • Implementation considerations: Back-EMF controllers must be evaluated for implementation on electronic processors or ASICs.The implementation choice is considered alongside the control method’s practical requirements.
  • Open-loop starting: All back-EMF sensing methods require an open-loop starting technique because back-EMF-based position information is unavailable at standstill.The paper treats open-loop starting as a shared implementation issue across these methods.

5.1. Comparison of methods’ feasibility

Back-EMF sensing methods exploit the unexcited phase of six-step BLDC commutation, but their practical feasibility differs across speed ranges and noise conditions. Third-harmonic integration extends the usable range relative to conventional approaches, while low-speed operation remains constrained by weak or absent back-EMF.

  • Back-EMF sensing basis: During 120° six-step conduction, two phases carry current and the third phase is available for back-EMF sensing.This unused phase provides the electrical signal needed by common sensorless methods.
  • Back-EMF sensing methods: Terminal-voltage zero-crossing sensing avoids measuring motor neutral voltage, while integration methods reduce sensitivity to switching noise.The integration approaches still face low-speed accuracy problems, and diode-state sensing is comparatively complicated and costly.
  • Third-harmonic sensing: Third-harmonic sensing offers a wider speed range and smaller phase delay than terminal-voltage sensing, but accumulated noise and offset can cause low-speed position errors.Lower signal levels also make third-harmonic and phase-voltage zero-crossing detection difficult at low speeds.
  • Comparison of methods: 6,000 rpm to about 100 rpm is the reported operating range for third-harmonic control, compared with about 6,000 rpm to 1,000 rpm for conventional back-EMF control.The third-harmonic method requires less filtering and introduces less phase delay than the zero-crossing method.
  • Low-speed feasibility: Back-EMF sensing cannot operate effectively at standstill or low speed because the back-EMF is zero or undetectably small, so open-loop starting is required.This low-speed limitation applies across back-EMF-based sensorless techniques.

5.2. Open-Loop Starting

Open-loop starting supplies a gradually advancing stator field until the rotor begins rotating and back-EMF sensing can be used. The approach is simple but its reliability and permissible behavior depend on load, switching speed, and application requirements.

  • Starting principle: Open-loop starting gradually increases the magnitude and/or frequency of a rotating stator field from standstill.The rotor turns once the stator field overcomes friction and inertia sufficiently.
  • Starting sequence: A representative procedure excites two phases for a preset interval, advances commutation by 120°, and then alters line-current polarity.The switching sequence establishes the initial rotor direction before continued commutation.
  • Starting limitations: The procedure is simple, but load conditions can reduce reliability and may cause temporary reverse rotor rotation during startup.Reverse motion is unacceptable for disk drives but may be satisfactory for pump and fan drives.
  • Starting limitations: Excessive stator-field speed when the rotor synchronizes can cause rotor oscillation during startup.The starting field must therefore be coordinated with the rotor’s ability to follow it.
  • Alternative starting methods: Stator-iron magnetic saturation can support initial rotor-position detection and speed-up methods such as short-pulse sensing.The short-pulse scheme uses successive short and long voltage pulses to generate positive torque.

5.3. Development of Controllers Using Electronic Processors and Specific Circuits

Electronic processors and dedicated circuits implement sensorless BLDC control through back-EMF sensing, PWM strategies, and integrated control hardware. These implementations reduce sensing hardware but introduce practical trade-offs involving filtering, commutation timing, noise, and high-speed operation.

  • Commercial processors and specific circuits support different BLDC control implementations, including back-EMF sensing techniques and processor-based control models.
  • DSPs enable complicated control algorithms at high sampling and calculation frequencies, while position sensors may not provide optimal performance under parameter variation.
  • Third-harmonic voltage processing can derive air-gap flux and rotor speed from the summed stator phase voltages.
  • PWM-based sensorless control for high-speed applications can suffer commutation delays because PWM switching and inverter commutation cannot always operate independently.
  • FPGA-based DVD spindle drives use PWM generation and power-device control while sampling motor terminal voltages for sensorless operation.
  • Single-terminal-voltage sensing and integrated filtering can reduce sensing-circuit component count and system cost, while microprocessors can derive speed from detected signal frequency.
  • Microcontrollers can extract true back-EMF zero crossings without a manufactured neutral voltage or extensive filtering, reducing total system cost.
  • ASIC-based control integrates the unenergized-winding terminal voltage and uses a PLL for commutation frequency and phase, but free-wheeling diode pulses can retard high-speed commutation.

6. Applications of BLDC Motor Controllers

BLDC controllers are applied across constant-load, varying-load, and positioning systems, with sensorless methods selected according to speed range, load dynamics, accuracy, and environmental constraints. Applications include appliances, automotive systems, aerospace actuators, hydraulic equipment, and high-speed storage devices.

  • BLDC controller applications are grouped into constant-load, varying-load, and positioning categories.
  • Constant-load applications: Constant-load applications prioritize variable speed over set-speed accuracy and commonly use low-cost controllers operating mostly in open loop.
  • Constant-load applications: Air-conditioner compressors demonstrate sensorless speed control over a wide range, with capacity modulated according to load for energy saving and comfort.
  • Constant-load applications: An IPM BLDC compressor drive operates from about 500 to 7,500 rpm and uses neutral-voltage variation to obtain rotor-position information when back-EMF amplitude is nearly zero.
  • Varying-load applications: Varying-load applications demand high-speed-control accuracy and good dynamic response, including washers, dryers, refrigerators, fuel pumps, steering, engines, and electric vehicles.
  • Varying-load applications: Returnless automotive fuel pumps use true back-EMF zero-crossing detection without manufactured neutral voltage or extensive filtering, providing a wider speed range from startup to full speed.
  • Varying-load applications: Aerospace EHA and EMA actuators require reliable BLDC control algorithms for safe startup and operation in safety-critical systems.
  • Positioning applications: HDD spindle motors face low back-EMF amplitudes, terminal-voltage spikes, reduced supply voltage, and difficult stable startup under mechanical disturbance.

7. Conclusions

The paper reviews BLDC position-control methods, emphasizing back-EMF schemes, estimators, implementation advances, and applications. It classifies existing and newer methods by their merits and drawbacks, highlighting sensorless control for cost, reliability, and harsh environments.

  • The paper reviews BLDC position-control methods, primarily back-EMF schemes and estimators, and discusses advances and applications.
  • The paper classifies existing and newer control methods together with their merits and drawbacks.
  • Eliminating shaft encoders, resolvers, or Hall-effect probes can further reduce cost and increase reliability.
  • Sensorless control is the only choice in some applications where position sensors cannot function reliably under harsh environmental conditions and higher performance is required.
Loading 2402.05263v1…