A motor controller is the electrical and software boundary between a power source, a motor and the mechanism that must move. It converts a motion command into switched voltage and current, interprets feedback, limits energy, handles faults and communicates what the drive is doing.
That definition is broader than a driver IC or an algorithm. A reliable controller is a system whose power stage, sensing, real-time firmware, thermal path, EMC behavior, interfaces and verification plan agree with the motor and load.
Requirements come before topology
Begin with operating points and abnormal conditions:
- continuous and peak torque, speed, acceleration and duty cycle;
- motor resistance, inductance, back EMF, pole pairs, inertia and feedback;
- DC-bus nominal, minimum, maximum, source impedance and transient range;
- startup load, stall, reversal, braking, back-driving and regeneration;
- speed, position, torque-ripple, vibration and acoustic limits;
- ambient temperature, cooling, contamination, moisture and cable length;
- communication, diagnostics, safety and production-test requirements.
Motor current cannot be selected from rated torque alone. Acceleration and disturbance rejection create peaks. The DC bus may sag when current rises and climb during regeneration. Wiring inductance adds switching stress. The controller should be specified for the complete mission profile.
Power conversion topologies
Brushed DC motor
A bidirectional brushed motor normally uses an H-bridge: two half-bridges steer voltage across the armature. PWM controls average voltage or current. Recirculation paths during drive, brake and coast affect loss, current measurement and bus energy.
Three-phase BLDC or PMSM
A three-phase permanent-magnet motor normally uses three half-bridges. Six power switches connect each phase to the positive or negative bus under controlled timing. Six-step, sinusoidal and field-oriented algorithms can use the same broad inverter topology while imposing different sensing and timing requirements.
Integrated versus external power devices
An integrated-FET driver can reduce size, parasitic inductance and design effort within its voltage, current and thermal limits. An external-FET gate driver gives the engineer more freedom to select MOSFETs, parallel devices and the thermal path. The trade is a larger, more layout-sensitive switching system.
Texas Instruments’ BLDC driver resources organize these choices around size, thermal efficiency, integration and control features. Device selection still needs a system loss and transient analysis.
Gate drive and switching behavior
The gate driver must charge and discharge each MOSFET gate at a controlled rate, translate logic-level commands to the high- and low-side devices, provide undervoltage behavior and prevent destructive cross-conduction. Important parameters include source/sink current, propagation delay, mismatch, dead time, bootstrap or charge-pump behavior, negative switch-node tolerance and gate protection.
Switching faster can reduce transition loss but increase voltage overshoot, ringing and electromagnetic emissions. Slower edges can reduce those effects while increasing switching loss. Gate resistance, driver strength, MOSFET charge, commutation loop inductance and PCB geometry determine the observed result. The oscilloscope measurement must use an appropriate probe and a short return; poor probing can create a misleading waveform.
TI’s three-phase gate-driver layout guidance emphasizes short phase and gate connections, careful sense routing and the relationship between MOSFET placement and protection measurements.
Current, voltage and position sensing
Current sensing
Current feedback supports torque regulation, current limiting, diagnostics and protection. A low-side bus shunt is compact but cannot observe every recirculation path. Phase-leg shunts can support phase-current reconstruction when PWM provides usable sample windows. Inline sensing observes phase current directly but has a demanding common-mode environment. Hall or fluxgate sensors can provide isolation at additional cost and size.
Shunt value trades signal amplitude against power loss. Amplifier offset, gain error, common-mode rejection, bandwidth, settling and recovery affect the usable measurement. Kelvin connections and PWM-synchronized ADC sampling matter. TI’s motor-drive current-sensing guide explains why control and protection depend on measurement latency and accuracy.
Voltage sensing
Bus voltage supports modulation limits, undervoltage/overvoltage response and regeneration handling. Phase voltage may support back-EMF detection or observers. Divider impedance, filtering, ADC range, switching transients and isolation must fit the algorithm.
Rotor feedback
Hall sensors provide electrical sectors. Encoders provide incremental or absolute angle. Resolvers tolerate harsh environments with additional interface electronics. Sensorless estimators infer position from electrical behavior, reducing hardware while adding startup, model and low-speed constraints.
Control architecture
Six-step commutation advances the stator field through six sectors per electrical revolution. It can use Hall sensors or sensorless back-EMF detection and may include closed-loop speed or current regulation. It is often attractive when modest processor load and simple commissioning matter.
FOC transforms measured current into a rotor-aligned coordinate frame and regulates torque- and flux-related components. It can provide smooth torque and a strong foundation for velocity or position control when current sensing, rotor angle, execution timing and tuning are adequate.
The detailed FOC versus six-step comparison explains when each architecture is appropriate. Control method and feedback method should be selected separately.
Firmware is a state machine, not only a control loop
The fast loop may execute current measurement, angle processing, control calculations and PWM updates. Around it, production firmware needs explicit states for initialization, calibration, idle, alignment, startup, run, braking, shutdown, fault and recovery.
Each transition needs entry conditions, timeout behavior and a defined torque response. A controller should not produce an uncontrolled pulse because an encoder is disconnected during startup or a watchdog restarts while the motor is spinning.
Useful firmware functions include:
- ADC offset and sensor plausibility checks;
- command limiting and rate limiting;
- bus-voltage-aware modulation and current limits;
- temperature derating;
- fault latching, event records and controlled retry;
- parameter versioning, integrity checks and safe defaults;
- communications timeout and supervisory interlocks;
- production calibration and end-of-line test support.
Protection layers
Fast hardware protection should address events that can damage the power stage before software reacts, such as severe overcurrent or shoot-through. Firmware handles slower limits, diagnostics, derating, fault classification and recovery. One threshold rarely serves both purposes.
Consider phase short, short to bus or ground, blocked rotor, open phase, loss of feedback, sensor saturation, bus transients, overtemperature, loss of gate-drive supply and reverse energy. Protection thresholds must include measurement tolerances and switching behavior so normal transients do not create nuisance trips.
Power source and regeneration
A motor controller is not a one-way load. During braking or back-driving, energy can return to the DC bus. A battery, bench supply and rectified AC supply respond differently. If the source cannot absorb energy, the bus needs another path or the motion profile must limit regeneration.
Bulk capacitance controls local bus impedance but should not be treated as an unlimited brake-energy store. Capacitance, ESR, ripple rating, wiring, precharge, inrush, discharge and fault energy belong in the power architecture.
Thermal and EMC design
Estimate conduction, switching, gate-drive, shunt, regulator and magnetics losses at representative operating points. Map the heat path through copper, vias, PCB, interface materials, enclosure and airflow. Validate the mission profile, including repeated acceleration and stall-limited holding torque.
EMC begins with current loops and edge rates. Minimize high-di/dt loop area, control switch-node coupling, separate power and measurement returns intentionally, provide local decoupling and design the motor-cable interface for the installation. Filters cannot rescue an uncontrolled layout without cost and side effects.
Interfaces and system integration
Define command source, update rate, latency, units, limits and loss-of-command behavior. Decide which parameters are fixed, calibrated or field-configurable. Diagnostic data should help distinguish motor, wiring, mechanics, sensor, power and controller faults.
Mechanical integration includes connector keying, strain relief, creepage/clearance where applicable, heatsink flatness, enclosure grounding, service access and assembly tolerance. Firmware and PCB work cannot be separated from those constraints.
Verification plan
- characterize the motor and load rather than relying only on nominal data;
- measure switching waveforms, gate behavior and current-sense timing;
- test startup, reversal, braking and regeneration at voltage/load/temperature corners;
- run continuous and transient thermal profiles;
- inject feedback, communication and power faults;
- evaluate conducted/radiated emissions and immunity with production cables;
- verify parameter, bootloader and interrupted-update behavior;
- build production tests around failures that matter.
What a complete controller design delivers
The deliverable is not a schematic plus firmware that spins a motor. It is a characterized drive with defined operating limits, controlled faults, validated thermal and EMC behavior, maintainable software, production tests and interfaces that fit the product.
Outer Reef designs custom motor-controller electronics and embedded firmware around the motor, load, power source, mechanical envelope and verification requirements.
