How Do You Drive a Brushless DC (BLDC) Motor with a Three-Phase Driver?

Detailed Explanation

A BLDC motor inverts the brushed DC motor's mechanical commutation (the brushes and commutator that physically switch winding current as the rotor turns) into an electronic one — the permanent magnets are on the rotor, the windings are fixed on the stator, and a driver circuit switches current through the windings in the correct sequence based on rotor position. This is more complex to drive than a brushed motor but eliminates brush wear, gives higher achievable efficiency and power density, and enables much more precise torque and speed control — which is why BLDC motors dominate applications from drone propulsion to appliance motors to electric vehicle traction where brushed motors previously served.

The Three-Phase Bridge

Where an H-bridge uses four switches to drive one brushed-motor winding in either direction, and a stepper driver uses eight switches (two H-bridges) for two windings, a BLDC three-phase bridge uses six switches — three half-bridges, one per winding (commonly labelled U, V, and W phases). At any commutation step, exactly two windings are actively driven (one sourcing current, one sinking it) while the third is either floating (trapezoidal commutation) or also actively driven at an intermediate voltage (sinusoidal/FOC commutation).

Commutation: Knowing When to Switch

The fundamental problem a BLDC driver solves is timing: switching the wrong windings, or switching at the wrong moment, produces reduced torque, increased current draw, or a stalled/reversed motor. Two approaches to knowing rotor position dominate:

Sensored (Hall-effect) commutation. Three Hall-effect sensors, positioned 120 electrical degrees apart inside the motor, each output a digital high/low signal indicating which magnetic pole is currently passing. The combination of all three sensor states directly maps to one of six commutation sectors — simple, reliable at any speed including standstill, but requires the extra sensor wiring (typically 5 wires beyond the 3 motor phases: sensor supply, ground, and three Hall outputs) and adds a component that can fail.

Sensorless (back-EMF) commutation. As covered in the FAQ above, the undriven phase's back-EMF zero-crossing substitutes for a Hall sensor, eliminating the extra wiring and sensor cost — the majority of small drone, RC, and appliance BLDC drivers use this approach — but it cannot detect rotor position reliably at zero or very low speed, requiring an open-loop startup ramp before closed-loop commutation can begin. This startup phase is a common source of failed starts under load, since the motor is being driven blind without any confirmation it's actually following the commanded ramp.

Trapezoidal vs Sinusoidal (FOC) Control

Beyond knowing rotor position, the driver chooses how to shape the phase currents:

• Trapezoidal (six-step) commutation switches between six discrete states per electrical revolution, matching the Hall-sensor sector boundaries directly — simple to implement in firmware or dedicated hardware, but produces torque ripple at each 60-degree commutation step (felt as vibration or heard as acoustic noise, particularly at low speed).
• Field-Oriented Control (FOC) treats the motor's three-phase currents as a rotating vector in a rotor-aligned reference frame, using the Clarke and Park transforms to convert measured phase currents into direct (torque-producing) and quadrature (flux) components that PI control loops regulate independently — producing smooth, continuously-varying phase currents rather than six discrete steps. FOC needs accurate continuous rotor angle (not just sector information), current sensing on at least two phases, and meaningfully more computation per control loop iteration, but delivers smoother torque, lower audible noise, and better efficiency, especially at low speed — the reason FOC has become the default for higher-performance BLDC applications (e-bikes, drones, robotics) despite its added complexity.

Driver ICs and Integration Level

Three-phase BLDC driver ICs span a range of integration levels, mirroring the choice already covered for H-bridge motor drivers and stepper driver ICs:

• Three-phase gate driver ICs (TI DRV8323-class, similar parts from onsemi and Infineon) provide just the high-side/low-side gate drive with bootstrap supplies and shoot-through protection for six external MOSFETs — see MOSFET gate driver IC and bootstrap design for the bootstrap-capacitor mechanics this extends to three phases. All commutation logic, sensing, and control algorithm run in the host MCU or a dedicated motor-control peripheral.
• Integrated BLDC controller/driver ICs combine the gate drivers with an onboard commutation state machine (sensored or sensorless), reducing firmware complexity at the cost of less control over the exact commutation algorithm — a reasonable trade for products that need a motor spinning at a commanded speed without custom torque-control requirements.
• Complete driver modules (including hobbyist/drone ESCs) package the full power stage, gate drivers, and commutation firmware behind a simple speed-command interface (commonly a PWM or digital protocol input) — the fastest path to a working motor, at the cost of firmware/algorithm flexibility.

Many modern MCUs also include dedicated motor-control peripherals (timers with complementary PWM outputs and dead-time insertion, ADC triggering synchronised to PWM edges for current sensing) specifically to support driving the discrete-MOSFET approach without dedicating the CPU to bit-perfect PWM timing.

Zeus Design designs BLDC motor-control hardware and firmware — from driver IC selection through FOC or trapezoidal control implementation — as part of complete product electronics development.

Design Considerations

• Plan the sensorless startup ramp deliberately if going sensorless. A fixed open-loop ramp profile that works for an unloaded motor on the bench can fail to start the same motor under its actual working load — validate startup under worst-case load and inertia, not just a free-spinning bench test.
• Verify Hall sensor phase alignment against the winding phase order for sensored designs. A Hall sensor sequence that doesn't match the actual winding commutation order causes the motor to run backwards, stall, or run with severely reduced torque — this is a wiring/configuration check to make during bring-up, not something firmware can fully compensate for after the fact.
• Size the bridge and driver for stall current the same way as any other motor drive. BLDC motors, like brushed and stepper motors, draw their highest current at low speed under load — apply the same stall-current sizing discipline covered in the H-bridge motor drive page rather than sizing only for the nameplate running current.
• Budget the control loop's computational load honestly before committing to FOC. FOC's Clarke/Park transforms, PI loops, and space-vector modulation calculation must complete within each PWM period — verify the target MCU's available cycles at the intended PWM frequency before committing to FOC over the simpler trapezoidal approach, particularly on lower-end Cortex-M0/M0+ parts without hardware floating-point or DSP instructions.

Common Mistakes

• Assuming sensorless commutation works identically to sensored at all speeds. The zero-speed/low-speed blind spot is a fundamental limitation of back-EMF sensing, not a tuning problem — a design that genuinely needs reliable starts under load at unpredictable initial positions should reconsider Hall sensors rather than trying to perfect a sensorless startup ramp indefinitely.
• Under-provisioning bootstrap capacitor refresh time in low-speed or stalled operation. The same high-side bootstrap-capacitor refresh problem covered for single-phase gate drivers applies across all three high-side switches in a BLDC bridge — a commutation pattern that holds one high-side switch on for too long at low PWM frequency or low speed can starve its bootstrap capacitor.
• Neglecting EMI from the fast-switching three-phase bridge. A BLDC bridge switches at least as aggressively as a single-phase H-bridge, across three phases simultaneously — apply the same switching-loop and layout discipline used for any high-current switching stage, and budget EMC pre-compliance testing time rather than discovering the problem at formal certification.
• Confusing electrical degrees with mechanical degrees on a multi-pole-pair motor. A BLDC motor with more than one magnetic pole pair completes multiple electrical commutation cycles per single mechanical revolution — commutation timing, Hall sensor spacing, and back-EMF zero-crossing intervals are all specified in electrical degrees, and conflating the two produces commutation timing errors that scale with the motor's pole-pair count.